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.
4 Methodology T his chapter describes the committeeâs approach to its task. The commit- teeâs initial step was to conduct a comprehensive search of the scientific literature to identify studies of long-term health outcomes in humans that might be associated with exposure to depleted uranium. The committee devel- oped criteria for evaluating the relevance and quality of studies. The selected studies constituted the primary evidence from which the committee drew con- clusions about the relationships between uranium and specific long-term health outcomes. The committee then ranked the strength of the relationships by using the five-category system presented at the end of the chapter. Information-Gathering Strategy The committee used a multistep process to identify scientific studies of the long-term health outcomes of exposure to uranium compounds, including depleted uranium. Twelve data sourcesâincluding PubMed, the National Techni- cal Information Service, and Toxicology Literature Online (TOXLINE)âwere searched for the key phrase depleted uranium and the Medical Subject Heading (MeSH) terms uranium and uranium compounds. Uranium-related studies identi- fied by the committee that wrote Gulf War and Health, Volume 1: Depleted Ura- nium, Pyridostigmine Bromide, Sarin, Vaccines (IOM, 2000b; hereafter referred to as Volume 1) were added to the current committeeâs reference database. Addi- tional studies were identified from the reference lists of technical reports, books, and other documents. 73
74 updated literature review of depleted uranium The searches generated about 3,500 titles and abstracts, which were exam- ined to identify articles that appeared to be relevant to the committeeâs task, that is, articles on health outcomes of exposure to uranium (including both natural and depleted uranium). Examples of types of articles that were excluded during this step are environmental studies (that is, effects on wildlife), engineering studies of nuclear reactors, studies of the treatment and disposal of uranium, studies of naturally occurring uranium concentrations in various locations, bioremediation studies, and studies of nuclear-plant safety. After examination of the titles and abstracts, about 1,000 articles remained in the committeeâs reference database. The data sources listed above were searched on a monthly basis through Decem- ber 2007 and relevant articles were added to the reference database. To gather further information, the committee held a public meeting on June 28, 2007, in Washington, DC. The topics discussed are shown in Box 4-1. Four speakers gave presentations related to health effects of exposure to depleted uranium and uranium in human populations, including the veteran population. Another presenter discussed toxicologic studies of uranium. After the formal presentations, the floor was opened to members of the public who wished to make comments. BOX 4-1 Open-Session Presentations, June 28, 2007 Depleted Uranium Exposure and Health Melissa McDiarmid, Department of Effects in Gulf War Veterans Veterans Affairs Depleted Uranium Follow-Up Program Outcomes of the UK Ministry of Nicholas Priest, Atomic Energy of DefenceâSponsored Depleted Uranium Canada Limited, Canada (formerly Research Programme with Middlesex University, UK) Depleted Uranium and Veteransâ Dan Fahey, Board Member, Veter- Health: A Flawed Testing Process and ans for Common Sense, and PhD an Undersized, Politicized Study Limit candidate, University of California, Evaluation of Exposures and Effects Berkeley Current NIOSH Research on Uranium- Mary Schubauer-Berigan, National Exposed Workers Institute for Occupational Safety and Health Toxic and Radiologic Effects of Fletcher Hahn, Lovelace Respira- U Â ranium: Animal Studies tory Research Institute Department of Defense Health Kenneth Cox, Department of Databases Defense
methodology 75 Principal Objectives of Epidemiologic Studies Epidemiologic studies examine the relationship between exposures to agents of interestâuranium in this caseâin a human population and the health out- comes seen in the population. The challenge of epidemiologic studies is to control for risk factors that are related to the exposures and health outcomes of interest by using study designs and statistical techniques that control for bias and confound- ing. Such studies can be used to generate hypotheses for future study or to test hypotheses posed by investigators. A principal objective of epidemiology is to understand whether exposures to specific agents are associated with disease or other health outcomes and to evalu- ate whether such associations are potentially causal. Although they are often used synonymously by the general public, the terms association and causation have distinct meanings (Alpert and Goldberg, 2007). Epidemiologic studies can establish statistical associations between expo- sures and health effects, and associations are generally expressed by using rela- tive risks or odds ratios. To conclude that an association exists, it is necessary for an exposure to be followed by a health effect more frequently than would be expected by chance alone. Furthermore, confidence in an association rises when it is consistently observed in several studies. However, the results of separate studies are sometimes conflicting. It is possible to attribute discordant study results to differences in such characteristics as soundness of study design, quality of execution, and the influence of different forms of bias. Studies that result in a tight confidence interval around a statistically significant relative risk of associa- tion constitute stronger evidence of an effect. When the measure of association does not show a statistically significant effect, it is important to consider the size of the sample and whether the study had the power to detect an effect of a given size. Epidemiologic study designs differ in their ability to provide valid estimates of an association (Ellwood, 1998). Cross-sectional studies generally provide a lower level of evidence than cohort and case-control studies. Determining whether a given statistical association rises to the level of cau- sation requires inference (Hill, 1965). As discussed by the International Agency for Research on Cancer (IARC) in the preamble of its monographs on evaluating cancer risks (for example, IARC, 2004), a strong association is demonstrated by repeated observations in a number of studies, an increased risk of disease with increasing exposure or a decline in risk after cessation of exposure, and specificity of an effect. Those characteristics all strengthen the likelihood that an association seen in epidemiologic studies is a causal effect. Inferences from epidemiologic studies, however, are often limited to population or ecologic associations because of a lack of information on individual exposures. Exposures are rarely, if ever, controlled in epidemiologic studies, and there is usually large uncertainty in the assessment of exposure. To assess whether explanations other than causality are responsible for an observed association, one must bring together evidence from
76 updated literature review of depleted uranium different studies and apply well-established criteria, which have been refined over more than a century (Hill, 1965; Susser, 1973, 1977, 1988, 1991; Evans, 1976; Wegman et al., 1997). For a review of those criteria, see the 2004 report of the US Surgeon General (2004). In examining the available epidemiologic studies, the committee addressed the question, âDoes the available evidence support a causal relationship or an association between exposure to uranium and a health effect?â Even a causal rela- tionship between exposure to uranium and a specific health effect would not mean that uranium invariably results in the health effect or that all cases of the effect are the result of uranium exposure. Such complete correspondence between exposure and disease is the exception in large populations (IOM, 1994b). The committee evaluated the data and based its conclusions on the strength and coherence of the data in the selected epidemiologic studies that met its inclusion criteria. Factors Influencing the Relevance and Quality of Studies The committee considered several important issues in its evaluation of the epidemiologic studies and assessment of evidence on uranium-processing work- ers and civilian and deployed populations exposed to natural and depleted ura- nium. The discussion builds on the topics covered in Volume 1 inasmuch as many of the methodologic issues are common to the old and new studies. Like the past committee, the present committee considered a number of factors in its evalua- tion, including measurement of exposure, assessment of outcome, and relevance of the study population to veterans. As a result, the committee has also outlined the common limitations in the studies on which it based its opinions. The basic limitations are a lack of representativeness or applicability, potential selection bias, incomplete control for potential confounders, and exposure and outcome misclassification. Limitations peculiar to specific types of studies are included in the discussion of those studies (see Chapter 7). Study Populations Relevance to Veteran Populations One of the most important potential limitations of existing studies is the lack of applicability to veteran populations. With the exception of the studies on deployed populations, many of the study samples differed from the Gulf War and Operation Iraqi Freedom troops with respect to characteristics and risk factors. The uranium- processing workers, for example, were mostly white men; few analyses exam- ined health effects in minority groups or women. Some of the residential studies included people of all ages (such as children and the elderly) living in a particular region with little regard to demographics, lifestyle, or other risk factors.
methodology 77 Comparison-Group Issues Many of the cohort studies of occupationally exposed workers described in Chapter 7 compared death rates in workers with death rates in the US popula- tion (or the population of the counties or states in which uranium workers lived). Those studies reported the standardized mortality ratio (SMR) because it is the principal means used in occupational studies to express the death rate in work- ers relative to that in people not exposed to the agent being studied. A statisti- cally significantly increased SMR (greater than 100 with 95% confidence limits that do not include 100) indicates the possibility of an association between an exposure and a disease. Replication of such a finding strengthens the evidence of an association. In the case of the occupational cohorts, the results might be skewed because of the âhealthy-worker effect.â That uranium processors were routinely subject to health examinations before and during employment means that this cohort was generally healthier than the population at large. Thus, one would expect mortality and morbidity in workers to be lower than those in the general popu- lation as reflected in study results regardless of exposure; this is known as the healthy-worker effect. It should be noted that studies of military populations are often subject to the same bias (the âhealthy-warrior effectâ) and this type of selection bias may result in underestimation of the association of the exposure and an outcome. The best way to avoid the healthy-worker bias is to use an internal com- parison group of employed people who did not receive the exposure of interest. However, even internal comparison groups can be subject to the bias to the extent that less healthy workers may not stay in more physically demanding jobs or jobsâlike uranium millingâthat may involve greater exposure to chemical agents. Studies that use internal comparison groups are generally more valid than studies that use external comparison groups (such as the US population or the population of a region) but are also subject to bias. It may be difficult to draw conclusions from studies that directly compare the SMRs of groups of workers that experienced different levels of radiation exposure, because the influence of confounding variables may differ between the groups. Exposure Assessment Methods for measuring exposure varied among the three study typesâ studies of uranium processors and of civilian and deployed populations. Studies of occupational exposure on which the committee relied heavily to evaluate the effect of uranium on disease used several methods and models to assess exposure, including direct measurement of individual exposure through estimates of inter- nal and external radiation dose, the use of work histories to estimate cumulative exposure, and classification of workers by maximum exposure.
78 updated literature review of depleted uranium Direct Measurement in Individual Workers The preferred method for an occupational study (or for any study) is to mea- sure exposure of each worker directly. Radiation film badges give a measure of cumulative exposure but respond only to external radiation, which is of greater concern for exposure to enriched uranium than to natural or depleted uranium. Measuring the internal dose of radiation is more difficult. The best method is mathematical modeling to infer the lung dose of uranium from measurements of uranium in the urine or ambient dust. However, the direct-measurement approach requires that a company moni- tor each individual worker for radiation exposure and keep thorough, accurate records. In many of the occupational retrospective cohort studies, the authors found that measurements of exposure in individual workers were either unavail- able or unreliable. In some cases, records were incomplete, so measurements missing for many workers were estimated from earlier periods or neighboring worksites. In other cases, the only measurements were of urinary uranium excre- tion. The body excretes uranium rapidly, so urinary uranium is a measure only of exposure in the preceding several days, not over the extended work period. Using Work History to Model Cumulative Exposure Several researchers approximated individual exposure by modeling cumula- tive exposure on the basis of a workerâs job history in the plant and the level of exposure in each worksite. They measured uranium exposure in various worksites in the processing plant, using measures of urinary uranium or uranium in ambi- ent dust. That information was used to model the cumulative lung dose per unit time in the worksite. They then used plant employment records to determine the amount of time that each worker spent in each job. By totaling each workerâs cumulative exposure in various worksites over the course of the workerâs period of employment, they estimated the workerâs total exposure. The modeling approach in effect assigns to each worker the average exposure in each worksite. Compared with individual direct measurement, this approach loses specific information because workers in a given site may vary in their exposure. Any approach that randomly misclassifies individual workersâ expo- sure levels while accurately tracking their health outcomes will result in muted estimates of association between exposure and outcome. That biases a study toward failing to detect an association between exposure and a health outcome even if one exists. Classifying Workers by Maximum Exposure This approach measures average exposure in each worksite, as described in the preceding section, and classifies worksites into a relatively small number
methodology 79 of groups according to the level of exposure. However, instead of estimating cumulative exposure over all worksites, this method uses as the exposure level for a worker the highest exposure level among all sites to which the worker was assigned for a minimum period (usually 1 month). This approach of exposure modeling is even cruder because it reduces the variation among workersâ exposure levels in two ways. First, it assumes that an employee spent his or her entire period of employment in one group of work- sites, whereas the worker may have spent time in sites that varied considerably in exposure levels. Second, it combines sites that may vary considerably in their level of exposure. For those reasons, this approach is especially prone to false- negative results (that is, failing to detect a dose-response relationship). However, the effect of the shortcomings of this approach is unknown because none of the studies estimated the probability of false-negative results. Self-Reporting in Exposure Assessment Self-reporting of exposure is a potential limitation. Peopleâs ability to recall details of exposure over a period of years accurately can vary widely and is likely to be small. In addition, recall can be influenced (that is, biased) by whether a person has experienced an adverse health outcome. Relying on memory can result in imprecise and even invalid assessments of exposure. Other Methods of Estimating Exposures A study that does not classify workers according to exposure cannot use workers with low exposure as an internal control when estimating the health effects of high exposure. Such a study must use the US population or the popu- lation of the region in which the plant is sited as the external control group. In that approach, the healthy-worker effect is more likely to distort estimates of the effect of exposure on health outcomes, generally biasing results toward lower risk among the exposed. Many of the studies were limited by potential exposure misclassification. Exposure can often be difficult to measure, particularly in settings where par- ticipants were exposed to a variety of compounds. For example, the uranium- processing workers received a variety of chemical and radiologic coexposures that are impossible to separate, so it is impossible to evaluate which exposure resulted in the outcome of interest. In addition, dose plays a crucial role in risk, and in many cases doses could not be determined. The studies also suffer from the lack of technical precision that is available today, so both overall exposure and dosage could be only imprecisely estimated on the basis of such surrogate factors as job classification or assignment. In the residential studies, except the drinking-water studies, exposure was generally not measured at the individual level, and exposure assessment was
80 updated literature review of depleted uranium based on geographic proximity modeling. In such studies, dose-response relation- ships cannot be determined. The same is true of the studies of deployed person- nel; in war situations (particularly combat, in which depleted-uranium exposure is most likely), environmental monitoring is not feasible. Outcome Assessment Biologic Plausibility Biologic plausibility reflects knowledge of the biologic mechanism by which an agent can lead to a health outcome. That knowledge comes through mechanism- of-action or other studies in pharmacology, toxicology, microbiology, physiology, and other fieldsâtypically in studies of animals. Biologic plausibility is often dif- ficult to establish or may not be known when an association is first documented. The committee considered such factors as evidence from animal and human studies that exposure to an agent is associated with diseases known to have bio- logic mechanisms similar to that of the disease in question, evidence that some outcomes are commonly associated with occupational or environmental exposures, and knowledge of routes of exposure, storage in the body, and excretion that sug- gests that a disease is more likely to occur in some organs than in others. The extent to which the data are consistent with a biologically plausible mechanism influences the weight attached to the results of a study, as does an indication that the mechanism is similar in the animals under study and humans. Biomarkers A biomarker is a molecular or cellular indicator of exposure, effect, or sus- ceptibility. More specifically, a biomarker of effect is defined as a âmeasurable biochemical, physiologic, behavioral, or other alteration in an organism that, depending on the magnitude, can be recognized as associated with an established or possible health impairment or diseaseâ (NRC, 2006). Biomarkers of effect can include biochemical, cellular, and physiologic indicators of disease. Numerous studies evaluated by the committee incorporated biomarkers of effect to evaluate health outcomes related to uranium exposure, including biomarkers to evaluate cell toxicity and renal dysfunction. Adequate Followup Period To strengthen the evidence of a true cause-effect association (particularly for some health outcomes, such as most cancers), the followup period should allow suffi- cient time after exposure for the health outcome to occur in the population of concern. There are several time-related factors. Biologic latency of cancer is a factor in the delay between exposure to a putative carcinogen and the appearance of cancer. For
methodology 81 most cancers, the lag between exposure and diagnosis is at least 10 years; however, there are exceptions, such as leukemia. Eliminating study participants who died from cancer that occurred within 10 years of exposure should increase the SMR if there is a true association between exposure to the agent and the cancer. Conversely, the case for an association is much weaker when the death rate relative to that in the US population is the same whether or not the author considered the early cancer deaths. Specificity of Outcome The study had to specify a distinct outcome rather than a nonspecific group of health outcomes. Lack of specificity occurs primarily in mortality studies that examine all-cause mortality (such as deaths from all types of cancer) as opposed to cause-specific mortality (such as deaths from lung cancer). All-cause mortality studies were excluded unless they analyzed specific health outcomes. Adverse Clinical Outcomes After reviewing the approximately 1,000 articles in the reference database, the committee focused on a number of relevant health outcomes on which to draw conclusions (see Chapters 6 and 8). The selected health outcomes are 10 types of cancer and several nonmalignant diseases or conditions. The types of cancer are lung cancer, leukemia, lymphoma, bone cancer, renal cancer, bladder cancer, brain and other central nervous system cancers, stomach cancer, prostatic cancer, and testicular cancer; the nonmalignant diseases or conditions include renal disease, respiratory disease, neurologic disease, and reproductive and devel- opmental effects. With the exception of prostatic and testicular cancers, the health outcomes were selected by the committee because there are plausible mechanisms of action (for example, lung cancer and respiratory disease were selected because inhaled insoluble uranium oxides lodge in the lung). Prostatic cancer is the most frequently diagnosed cancer in men in the United States, and any slight increase in risk could result in large numbers of cases and deaths. Testicular cancer, the most common cancer in young men, is of special interest to Gulf War veterans, and some recent studies of veterans suggested a higher but nonsignificant risk in them than in their nondeployed counterparts (IOM, 2006). Considerations in Statistical Inference Tests of Association Studies of a possible relationship between an exposure of interest and an outcome typically report statistical tests of association. Those tests assess whether the data are consistent with the claim of an association between exposure and outcome. The association is commonly expressed in terms of null and alternative
82 updated literature review of depleted uranium hypotheses. The null is chosen to be consistent with the âstatus quoââstatistical independence or no association between exposure and outcome. The alternative is chosen to represent the opposite point of view: that an association exists between exposure and outcome (that is, they are not independent). A summary statistic, called a test statistic, is calculated that gauges how well the data âmatchâ the null hypothesis. In general, small values of the test statistic reflect consistency with the null, and large values consistency with the alternative. The magnitude of the test statistic is compared with its expected size under the null hypothesis. The difference between the observed value of the statistic and its expected value under the null is evaluated while taking into consideration such factors as the size of the sample and variability of the measurements. The p Value In reporting the results of a statistical test of association, researchers report a p value, or the probability of observing a test statistic as large as or larger than (in absolute value) that obtained from the sample if the null hypothesis is true. Small p values therefore indicate that the probability of observing a result as extreme as or more extreme than that obtained in the study is very unlikely if the null is true. By convention, most researchers use a p value of 0.05 as the threshold value for rejecting the null hypothesis. Therefore, if researchers observe p < 0.05, they state that a result is âstatistically significantâ; if the p value exceeds 5%, they state that the result is nonsignificant. Type I and Type II Error and Power It is possible to make two types of errors in conducting a statistical test of association. First, the null hypothesis might be rejected when it is true, simply because of chance variation. That is called a type I error, or Î±. The second type of error is failure to reject the null hypothesis when the alternative is true. That is called a type II error. One minus the type II error, or the probability of reject- ing the null when the alternative is true, is called the power of a test. In general, researchers want both error rates to be low. In practice, the type I error is usu- ally set to an acceptable level (usually 5%, as indicated above), and a study is designed to obtain a suitably large value for the power. Power is a function of the size of the study sample, the duration of followup, and the strength of the exposure effect. Longer followup will also allow examination of a range of latent periods between exposure and diagnosis of disease. Control of Bias Bias refers to systematic or nonrandom error. Bias causes an observed value to deviate from the true value. It can weaken an association or generate a spurious
methodology 83 association. Because all studies are susceptible to bias, a goal is to minimize bias or to adjust the observed value of an association by using special methods to cor- rect for bias. Two kinds of bias may compromise the results of an investigation: selection bias and information bias. â¢ Selection bias occurs when the participants in a study are not representa- tive of the general population. The study participants differ from nonparticipants in characteristics that cannot be observed, that is, groups differ in measured or unmeasured baseline characteristics because of how participants were selected or assigned. â¢ Information bias results from the manner in which data are collected and can result in measurement errors, imprecise measurement, and misdiagnosis. Those types of errors may be uniform in an entire study population or may affect some parts of the population more than others. Bias may result from misclas- sification of study subjects with respect to the outcome variable. Other common sources of information bias are the inability of study subjects to recall accurately the circumstances of their exposure (recall bias) and the likelihood that one group more frequently reports what it remembers than another group (reporting bias). Information bias is especially harmful in interpreting study results when it affects one comparison group more than another. Coexposures Confounders Many of the studies reviewed failed to control for potential confounders. For many of the outcomes of interest, there are several well-known risk factors that were not taken into consideration; these include smoking, diet, other lifestyle factors, and preexisting illness. In some studies, the lack of control was a result of the study design; for example, ecologic studies, such as the residential studies in which exposure is determined solely by geographic proximity to an exposure source, by design cannot take individual-level factors into account. Retrospec- tive cohort studies can only be analyzed on the basis of the data available; often, information on other risk factors was not collected, either because they were not known risk factors at the time or because collection of such information was not routine. Synergism Interaction, or synergism, occurs when combined exposure to two or more chemicals is more likely to produce an adverse health outcome than exposure to the chemicals individually. Epidemiologic studies are typically unable to partition data on exposures to multiple chemicals quantitatively and even less likely to be
84 updated literature review of depleted uranium able to attribute health outcomes related to combined exposures. As discussed previously, many of the studies evaluated by the committee, including those with an extensive exposure-assessment component (that is, occupational studies), might not have been able to account for the numerous chemical or radiologic coexposures. Although the committee was not charged with evaluating health effects related to combined exposures to chemicals, it acknowledges the possibil- ity that such exposures are common, particularly in occupational settings. Epidemiologic-Study Designs The major types of epidemiologic studies evaluated by the committee are cohort, case-control, cross-sectional, ecologic studies, and case reports and case series. Cohort Studies A cohort, or longitudinal, study follows a defined group, or cohort, over time. It can test hypotheses about whether an exposure to a specific agent is related to the development of a health effect and can examine multiple health effects that may be associated with exposure to a given agent. A cohort study starts by classifying study participants according to whether they have been exposed to the agent under study, in this case uranium. A cohort study com- pares health effects in people who have been exposed with those in people who have not been exposed. Such a comparison can be used to estimate a risk difference or a relative risk, two statistics that measure association. The risk difference is the rate of disease or health effect in exposed persons minus the rate in nonexposed persons; a value greater than zero implies that an excessive rate of disease is associated with the exposure. The relative risk, or risk ratio, is determined by dividing the rate of the disease in the exposed group by the rate in the nonexposed group; a relative risk greater than 1 suggests a positive association between the agent and the health effect, and the higher the relative risk, the stronger the association. One major advantage of a cohort study is the ability of the investigator to define the exposure classification of subjects at the beginning of the study. The classification in prospective cohort studies (see below) is not influenced by the presence of a health effect, because the health effect has yet to occur, and this reduces an important source of potential bias known as selection bias. As explained in the next section, on case-control studies, when it is possible to measure a confounding factor, the investigator can apply statistical methods to A potential confounding factor is a variable that is associated with the health outcome and may affect the results of the study because it is distributed differently in the exposed and nonexposed groups.
methodology 85 minimize its influence on the results. Another advantage of a cohort study is that it is possible to calculate absolute rates of disease incidence. A final advantage, especially over cross-sectional studies (discussed below), is that it may be pos- sible to adjust each subjectâs followup health status in light of baseline health status so that the person acts as his or her own control instead of defining a group as âdisease-freeâ; this may reduce a source of variation and increase the power to detect effects. The disadvantages of cohort studies are the high costs associated with using a large study population and long periods of followup (especially if the health effect is rare), attrition of study subjects, and delay in obtaining results. A prospective cohort study selects subjects on the basis of exposure (or lack of it) and follows the cohort to determine the rate at which the health effect develops. A retrospective (or historical) cohort study differs from a prospective study in temporal direction; the investigator traces back in time to classify past exposures in the cohort and then tracks the cohort forward in time to ascertain the rate of the health effect. Retrospective cohort studies often focus on mortality because of the relative ease of determining the vital status of individuals and the availability of death certificates to determine the causes of deaths. Most cohort studies are retrospective. For comparison purposes, cohort studies often use general population mor- tality or morbidity rates (age-, sex-, race-, time-, and cause-specific) because it may be difficult to identify a suitable control group of nonexposed people. The observed number of deaths or cases of illness in a group (related to a specific cause, such as lung cancer) is compared with the expected number of deaths or cases of illness. The ratio of observed to expected deaths is an SMR. An SMR greater than 1.0 generally suggests an increased risk of deaths in the exposed group. The major problem in using general population rates for comparison with military-cohort rates is the healthy-warrior effect, which arises when a military population experiences a lower mortality or morbidity rate than the general population, a mixture of healthy and unhealthy people. The military has physical- health criteria that personnel must meet when they enter the military and while they are on active duty. Case-Control Studies In a case-control study, subjects (cases) are selected on the basis of having a health effect; controls are selected on the basis of not having the health effect. Cases and controls are asked about their exposures to specific agents. Cases and controls can be matched with regard to such characteristics as age, sex, and socioeconomic status to eliminate those characteristics as causes of observed Incidence is the rate of occurrence of new cases of an illness or disease in a given population during a specified period.
86 updated literature review of depleted uranium differences, or those variables can be controlled for in the analysis. The odds of exposure to the agent among cases are then compared with the odds of exposure among controls. The comparison generates an odds ratio, which is a statistic that depicts the odds that those exposed to the agent in question will have a health effect relative to the odds that those not exposed will have the health effect. An odds ratio greater than 1 indicates that there is a potential association between exposure to the agent and the health effect; the greater the odds ratio, the stronger the association. Case-control studies are useful for testing hypotheses about the relationships between exposure to specific agents and a health effect. They are especially useful and efficient for studying the etiology of rare effects. Case-control studies have the advantages of ease, speed, and relatively low cost. They are also valuable for their ability to probe multiple exposures or risk factors. However, case- control studies are vulnerable to several types of bias, such as recall bias, which can dilute or enhance associations between exposure and a health effect. Other problems include identifying representative groups of cases, choosing suitable controls, and collecting comparable information about exposures of cases and controls. Those problems might lead to unidentified confounding variables that differentially influence the selection of cases or control subjects or the detection of exposure. For the reasons discussed above, case-control studies are often the first approach to testing a hypothesis about whether factors contribute to a specific health effect, especially a rare one. A ânestedâ case-control study draws cases and controls from a previously defined cohort. Thus, it is said to be nested in a cohort study. Baseline data are collected at the time that the cohort is identified, and this ensures a more uniform set of data on cases and controls. Members of the cohort who are identified as having a health effect serve as cases, and a sample of those who are effect-free serve as controls. Baseline data are used to compare exposure in cases and con- trols, as in a regular case-control study. Nested case-control studies are efficient with respect to the time and cost needed to reconstruct exposure histories of cases and of only a sample of controls rather than the entire cohort. In addition, because the cases and controls come from the same previously established cohort, concerns about unmeasured confounders and selection bias are decreased. Cross-Sectional Studies The main differentiating feature of a cross-sectional study is that exposure information and health-effect information are collected at the same time. The selection of people for the studyâunlike selection for cohort and case-control studiesâis independent both of the exposure to the agent in question and of health-effect characteristics. Cross-sectional studies seek to uncover potential associations between exposure to a specific agent and development of a health effect. In a cross-sectional study, effect size is measured as relative risk, preva-
methodology 87 lence ratio, or prevalence odds ratio. The study might compare health-effect or symptom rates in groups with and without exposure to uranium. Cross-sectional studies are easier and less expensive to perform than cohort studies and can identify the prevalence of health effects and exposures in a defined population. They are useful for generating hypotheses, but they are much less useful for determining cause-effect relationships (Monson, 1990). It might also be difficult to determine the temporal sequence of exposures and symptoms or effect. Ecologic Studies Ecologic studies examine the relationship between exposure and disease in groups of people rather than individuals. They can be used as a first step to explore whether an association between exposure and disease exists or to iden- tify avenues of research to investigate etiologic relationships. Ecologic studies require aggregate data on disease and exposure. Data on disease occurrence are commonly derived from incidence and mortality data, and exposure information is often based on an overall index, for example, environmental data, such as air or water quality. Despite the advantages of being inexpensive and relatively less time- consuming, ecologic studies have numerous methodologic problems. Ecologic studies use population-level data rather than individual-level data to assess the relationship between exposure and outcome, so an observed association cannot be used to draw inferences at the individual level (what would be referred to as the ecologic fallacy). Other methodologic problems include difficulty in controlling for confounders, within-group misclassification, and an inability to detect complicated relationships because of the limited data used in the analysis. Case Reports and Case Series A case report is a detailed descriptive study of an individual (case report) or small group (case series) in which an association between a health effect and a specific exposure (in this case uranium) is evaluated on the basis of medical histories and clinical evaluations. Case reports are observational, can be prospec- tive or retrospective, and are most useful in assessing rare diseases or providing an indication that an adverse health effect is related to an exposure. Case-series studies also have a number of limitations: they are vulnerable to bias because the observations are generally uncontrolled and collected in an unsystematic man- ner, and the small number of observations makes it impossible to generalize the findings to a larger population. Prevalence is the number of cases of an illness or disease in a given population at a specific point or in a specific period.
88 updated literature review of depleted uranium Inclusion Criteria After securing the full text of the roughly 1,000 articles mentioned above, the committee had to determine which ones to include in its review. For a study to be included in the committeeâs review, it had to meet these criteria: â¢ It had to be a published in a peer-reviewed journal or have undergone an equally rigorous process. â¢ It had to include details of its methodology. â¢ It had to include a control or reference group. â¢ It had to use reasonable methods to control for confounders and minimize selection bias. â¢ It had to use appropriate assessment of uranium exposure in the popula- tion. It had to measure exposure to uranium separately from other exposures; studies that dealt with multiple exposures and did not specifically report uranium exposures were excluded. â¢ It had to deal with long-term health outcomes. î It had to have a followup time adequate to detect an effect. î It had to have appropriate outcome assessments and measurements based on the expected biologic mechanisms. â¢ It had to include a relevant study population. Relevant study populations are î Uranium-exposed workers (that is, uranium-processing workers). î Military personnel deployed to the Gulf War. î People living close to uranium-processing facilities. Uranium exposure of such people may be similar to the âlevel IIIâ exposure received by military personnel (see Chapter 5). Rationale For Not Including Studies of Uranium Miners The committee engaged in extensive discussion about the role of epidemio- logic studies of uranium miners in shaping its report and its conclusions. The committee that wrote Volume 1 âexamined studies of health effects in uranium miners, but concluded that these studies have limited relevance because the primary disease-causing exposures were not to uranium but to radon decay prod- uctsâ (IOM, 2000b). The current committee also concluded that epidemiologic studies in uranium miners have limited relevance, but for different reasons. As detailed elsewhere in this report, the committee attempted to separate radiologic and chemical effects. The main health effect of interest in uranium miners is lung cancer, which this committee deems a radiologic effect, although a chemical toxic effect cannot be ruled out. If it is a radiologic effect, it should be theoretically possible to convert exposures to mine-based uranium and battlefield-
methodology 89 based depleted uranium into equivalent radiation doses (including corrections for radiation quality) by dose reconstruction and to transfer radiogenic lung-cancer risk estimates from the uranium miners to Gulf War veterans. However, the com- mittee identified four issues related to confounding that substantially limited the usefulness of the uranium-miner studies: â¢ Differences in physicochemical properties. The uranium miners were exposed primarily to radon decay products, whereas the veterans were exposed to aerosolized uranium particles. The two kinds of material probably have differences in physicochemical properties, such as aerodynamic behavior and solubility, that could result in large differences in environmental distribution in different media and thus differences in exposure, in transfer from the environ- ment to humans, and in the amount and pattern of deposition in humans (internal dose). â¢ Deposition pattern in the human body, especially the airways. The ura- nium miners were exposed primarily to radon decay products that were attached to (adsorbed onto the surface of) other particulate matter in the mine air. That particulate matter was mostly large or âcoarseâ (greater than 1 Î¼m in diameter). Furthermore, the miners typically used a combination of nose and mouth breath- ing at a relatively high inspiratory flow rate. Those considerations resulted in aerodynamic properties that favored deposition in large, central airways. In con- trast, veterans were exposed to a wider range of particle sizes, and the toxicants of interest (depleted-uranium particles) were not necessarily attached to other particulate matter. Smaller particles would probably penetrate all the way to the periphery of the lung and result in exposure of different types of cells and in the possibility of different types of lung cancer from what was typically the case in uranium miners. â¢ Radiological vs chemical mechanisms of toxicity. If the mechanism of lung-cancer production in the uranium miners is radiologic and if the same and only the same mechanism applies to the veterans, the type of dose reconstruction and risk transfer identified above would be appropriate and useful. However, because depleted-uranium toxicity involves an unknown combination of radio- logic and chemical mechanisms, including the possibility of synergism between mechanisms, it is scientifically inappropriate to use the uranium-miner data in this fashion. â¢ Coexposures to other agents. As pointed out in Volume 1, the uranium âminers were exposed to other possibly toxic dusts and, potentially, to diesel gas fumes, which might cause cancer and other diseases of the lungâ (IOM, 2000b). Veterans were also potentially exposed to a suite of toxic dusts, including diesel gas fumes, and that suite probably differed from that to which the uranium min- ers were exposed. In addition, smoking, which could have differed in pervasive- ness and magnitude between miners and veterans, is an important confounding v Â ariable.
90 updated literature review of depleted uranium In light of those four issues, the committee followed the approach of the first committee: the uranium-miner data were studied but played only a small role in the final assessment of risk. Categories of Strength of Association The committee used the evidence in the scientific literature to draw conclu- sions about exposure to depleted uranium and specific adverse health outcomes. Its conclusions are presented as categories of strength of association. Origin of the Categories IARC, part of the World Health Organization, established criteria in 1971 to evaluate the human carcinogenic risk posed by chemicals (IARC, 1998). First published in 1972, IARCâs evaluations are scientific, qualitative judgments of ad hoc working groups about the evidence of carcinogenicity provided by the avail- able data. The working groups expressed their qualitative judgments by choos- ing from among five categories to describe the relative strength of the evidence that a substance or exposure is carcinogenic (IARC, 2006a). That agencies in 57 countries use IARCâs published evaluations reflects the wide acceptance of the categorization scheme as it has been updated and applied to about 900 agents, mixtures, and exposures (IARC, 2005, 2006b). In the early 1990s, an Institute of Medicine committee adopted IARCâs categories for its evaluation of the adverse health effects of pertussis and rubella vaccines (IOM, 1991). Later committees used the categories, with some modifications, in their evaluations of the safety of childhood vaccines (IOM, 1994a), of the health effects of herbicides used in the Vietnam War (IOM, 1994b, 1996, 1999, 2001, 2003), and of the relationship between exposure to indoor pollutants and asthma (IOM, 2000a). The categories also were adapted and used by the present committeeâs predecessors, which evaluated the health effects of vaccines given to US troops and of chemical, bio- logic, and physical exposures that may have occurred during the Gulf War (IOM, 2000a,b, 2004, 2005, 2007). The five categories used in this report are defined below, and the committeeâs conclusions are presented in Chapter 8. Sufficient Evidence of a Causal Relationship Evidence is sufficient to conclude that a causal relationship exists between the exposure to uranium and a specific health outcome in humans. The evidence fulfills the criteria for sufficient evidence of an association (below) and satisfies several of the criteria used to assess causality: strength of association, dose- response relationship, consistency of association, temporal relationship, specific- ity of association, and biological plausibility.
methodology 91 Sufficient Evidence of an Association Evidence is sufficient to conclude that there is an association. That is, a con- sistent association unlikely to be due to sampling variability has been observed between exposure to uranium and a specific health outcome in human studies that were free of severe bias and that controlled for confounding. Limited/Suggestive Evidence of an Association Evidence is suggestive of an association between exposure to uranium and a specific health outcome, but the body of evidence is limited by insuffi- cient avoidance of bias, insufficient control for confounding, or large sampling v Â ariability. Inadequate/Insufficient Evidence to Determine Whether an Association Exists Evidence is of insufficient quantity, quality, or consistency to permit a con- clusion regarding the existence of an association between exposure to uranium and a specific health outcome in humans. Limited/Suggestive Evidence of No Association Evidence is consistent in not showing an association between exposure to uranium of any magnitude and a specific health outcome. A conclusion of no association is inevitably limited to the conditions, magnitudes of exposure, and length of observation in the available studies. References Alpert, J. S., and R. J. Goldberg. 2007. Dear patient: Association is not synonymous with causality. American Journal of Medicine 120(8):649-650. Ellwood, J. M. 1998. Critical appraisal of epidemiological studies and clinical trials. 2nd ed. Oxford, UK: Oxford University Press. Evans, A. S. 1976. Causation and disease: The Henle-Koch postulates revisited. Yale Journal of Biol- ogy and Medicine 49(2):175-195. Hill, A. B. 1965. The environment and disease: Association or causation? Proceedings of the Royal Society of Medicine 58:295-300. IARC (International Agency for Research on Cancer). 1998. Preamble to the IARC monographs. http://www-cie.iarc.fr/monoeval/background.html (accessed July 19, 2005). âââ. 2004. Tobacco smoke and involuntary smoking. IARC monographs on the evaluation of carcinogenic risks to humans. Lyon, France: International Agency for Research on Cancer. âââ. 2005. Monographs on the evaluation of carcinogenic risks to humans. http://www-cie.iarc. fr/ (accessed August 3, 2005). âââ. 2006a. Evaluation and rationale. Preamble to the IARC monographs. http://monographs.iarc. fr/ENG/Preamble/currentb6evalrationale0706.php (accessed June 10, 2008).
92 updated literature review of depleted uranium âââ. 2006b. Objective and scope. Preamble to the IARC monographs. http://monographs.iarc. fr/ENG/Preamble/currenta2objective0706.php (accessed June 10, 2008). IOM (Institute of Medicine). 1991. Adverse effects of pertussis and rubella vaccines. Washington, DC: National Academy Press. âââ. 1994a. Adverse events associated with childhood vaccines: Evidence bearing on causality. Washington, DC: National Academy Press. âââ. 1994b. Veterans and Agent Orange: Health effects of herbicides used in Vietnam. Washing- ton, DC: National Academy Press. âââ. 1996. Veterans and Agent Orange: Update 1996. Washington, DC: National Academy Press. âââ. 1999. Veterans and Agent Orange: Update 1998. Washington, DC: National Academy Press. âââ. 2000a. Clearing the air: Asthma and indoor air exposures. Washington, DC: National Academy Press. âââ. 2000b. Gulf War and health, volume 1: Depleted uranium, sarin, pyridostigmine bromide, vaccines. Washington, DC: National Academy Press. âââ. 2001. Veterans and Agent Orange: Update 2000. Washington, DC: National Academy Press. âââ. 2003. Veterans and Agent Orange: Update 2002. Washington, DC: The National Academies Press. âââ. 2004. Gulf War and health: Updated literature review of sarin. Washington, DC: The Na- tional Academies Press. âââ. 2005. Gulf War and health, volume 3: Fuels, combustion products, and propellants. Wash- ington, DC: The National Academies Press. âââ. 2006. Gulf War and health, volume 4: Health effects of serving in the Gulf War. Washington, DC: The National Academies Press. âââ. 2007. Gulf War and health, volume 5: Infectious diseases. Washington, DC: The National Academies Press. Monson, R. 1990. Occupational epidemiology. 2nd ed. Boca Raton, FL: CRC Press. NRC (National Research Council). 2006. Human biomonitoring for environmental chemicals. Wash- ington, DC: The National Academies Press. Susser, M. 1973. Casual thinking in the health sciences: Concepts and strategies of epidemiology. New York: Oxford University Press. âââ. 1977. Judgment and causal inference: Criteria in epidemiologic studies. American Journal of Epidemiology 105(1):1-15. âââ. 1988. Falsification, verification, and causal inference in epidemiology: Reconsideration in the light of Sir Karl Popperâs philosophy. In Causal inference, edited by K. J. Rothman. Chestnut Hill, MA: Epidemiology Resources. Pp. 33-58. âââ. 1991. What is a cause and how do we know one? A grammar for pragmatic epidemiology. American Journal of Epidemiology 133(7):635-648. U.S. Surgeon General. 2004. The health consequences of smoking: A report of the surgeon general. http://www.surgeongeneral.gov/library/smokingconsequences (accessed October 26, 2004). Wegman, D. H., N. F. Woods, and J. C. Bailar. 1997. Invited commentary: How would we know a Gulf War syndrome if we saw one? American Journal of Epidemiology 146(9):704-711.