Cover Image


View/Hide Left Panel
Click for next page ( 298

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement

Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 297
28 Lead as a Paradigm for the Stucly of Neurodevelopmental Toxicity Because lead is the best-studied neuro- toxin, it is a useful paradigm through which to understand the problems encoun- tered in the search for valid and effective markers of exposure, effect, and vulnera- bility. A variety of markers of exposure to lead have been used over the last 3 decades. The utility, sensitivity, and specificity of these markers vary widely. Newer ana- lytic techniques that measure lead in smaller samples and at lower concentra- tions have permitted numerous epidemio- logic studies of effects to be conducted in many countries (Needleman, 1988~. This has resulted in sharper hypothesis for- mation. Also, the tissue concentration of lead considered to be toxic has been lowered. MARKERS OF EXPOSURE TO LEAD There have been many improvements in the biologic markers of exposure to lead. The following discussion includes the relatively imprecise use of surrogate markers and the more precise physiologic assessments. However, each marker has characteristics that limit its use and interpretation. 297 Surrogate Markers When direct sampling of body tissues is not possible, investigators use sur- rogates to measure burden, such as employ- ment history (e.g., length of employment in the lead industry and type of job) and residence history (distance from a lead source-a smelter or a heavily traveled roadway). These surrogates for actual tissue concentrations are subject to large measurement errors, which reduce the spe- cificity (power) of a study. Blood Lead The most common marker of exposure is lead concentration in the blood. It has two weaknesses: Blood might not accurately reflect concentration in the tissue of interest, e.g., the brain or the testis. Blood lead content reflects only re- cent exposure (Rabinowitz et al., 1976~. Therefore, measurement of current con- centrations of blood lead could result in misclassification of subjects whose exposure ended before the sampling. By blurring the distinction between exposed and unexposed groups or amplifying the

OCR for page 297
298 error in measuring critical tissue con- centrations, these weaknesses in the bio- logic marker can increase the false-nega- tive rate of a study. Hair Lead Hair seems to be an ideal tissue by which to index body burden. Not only is keratin an avid sink for lead, but the tis- sue is accessible, and its collection is painless. It is difficult, however, to determine whether the lead measured in an assay of hair is externally deposited or represents internal dose. The concen- tration of lead in hair is also sensitive to the type and frequency of shampooing (Limic and Valkovic, 1986~. Erythrocyte Protoporphyrin Lead interferes with heme production at several sites. It competes with iron in the presence of ferrochelatase and blocks the entry of iron into the heme poc- ket (Needleman, 1988~. Instead, zinc en- ters the heme pocket, and the entry results in an abnormal heme compound, zinc proto- porphyrin (ZPP). ZPP is easily measured and is correlated with internal lead burden (Piomelli et al., 1973~. It is used widely as a screening tool, because it is inexpen- sive and insensitive to contamination with lead on the skin. The assay does have the disadvantage of missing early lead expo- sures, and it might be insensitive to lead at internal doses below 30 ~g/dl. Because ZPP or its extracted compound-free eryth- rocyte protoporphyrin-is also increased In Iron deficiency, this assay for lead exposure can be confounded by the iron status of the subject; this lowers the specificity of the assay. ZPP measurement is one example of a biologic marker that can be considered a marker of either ex- posure or effect. Chelatable Lead The most critical measure of exposure is the amount of lead available to general circulation-also known as the chelatable- lead pool. The amount of chelatable lead NEURODEVELOPMEI~TAL TOXICOLOGY in children is determined by administering a calculated dose of edathamil calcium disodium (CaNa2EDTA), 500 mg/m2 of body surface area, and measuring the resulting excretion of lead in the urine over a stipu- lated period. Ratios of urinary lead to the EDTA dose are calculated; ratios great- er than 0.6 fig of lead to 1 me of EDTA over 8 hours or 1.0 fig of lead to 1 me of EDTA over 24 hours are considered evidence of excess lead storage and indications for thera- peutic chelation. Although the chelat- able-lead pool is the best index of meta- bolically available lead, its measurement requires close supervision, often in an inpatient setting. Bone Lead Lead is taken up by bone and is associated with the apatite crystal, much as calcium is (Schwartz et al., 1988; Silbergeld et al., 1988~. In children with long, intense exposure, x-ray pictures of the long bones can show increased densities at the meta- physes. That is thought to be due not to the presence of lead, but to increased calcification secondary to the effect of lead on osteoblast activity (Silbergeld et al., 1988~. Its specificity as a marker is good, because few conditions yield simi- lar x-ray findings. However, the sensitiv- ity of the assessment is low, and rather intense exposures to x rays are required to show the change (Silbergeld et al., 1988~. Tooth Lead Bone biopsy is an impractical measure, but lead in dentin of deciduous teeth is a good marker for integrated exposure over time (Altshuller et al., 1962; Needleman et al., 1972, 1974; Steenhout and Pourtois, 1981~. This tissue is correlated with blood lead content obtained 3 years before shedding. Subjects with increased dentin lead were found to have lower IQ scores, shorter attention spans, and greater speech and language impairment than con- trols, after adjustment for relevant so- cioeconomic and biologic covariates (Needlemanet al., 1979~.

OCR for page 297
LEAD ASH PARADIGM X-ray Fluorescence of Bone In viva x-ray fluorescence of bones has been used successfully to classify exposure in industrial settings (Ahlgren et al., 1980~. Also, x-ray fluorescence of in situ deciduous teeth has been repor- ted to be a useful marker (Bloch et al., 1977~. The exposure time in the trials was long, so the technique is difficult to use in screening studies of children. MARKERS OF EFFECT The definition of "adverse health ef- fect" is controversial. When a single molecule enters a cell, it binds to a ligand and thus alters the state of the cell. That might be considered an adverse health ef- fect. Considerable debate has surrounded the question of when a biochemical change becomes a health effect. Hernberg (1972) described a model that clarified the issue considerably. If lead dose were plotted on the abscissa and rela- tive frequency of given effect were plotted on the ordinate, measurement of a large number of outcomes would be expec- ted to produce a display like Figure 28- 1. Figure 28-2 displays the resulting relationship between intensity of effect and definition of toxicity. The dose- ~ on Cal z z O LLI CL ~ on C' llJ O: ~ ~ > v _ ~ no m g ~ LLI LL or: O 299 effect relationship is assumed to be monotonic, but could be linear, exponen- tial, or logarithmic. The display in Figure 28-2 takes into consideration the effect of individual judgment on definition of toxicity. For noncritical, non-rate-limiting systems, there would be some effect with small in- tensity that all observers would agree is not an adverse health effect; that is, the effect bears no measurable relation- ship to longevity, reproductive efficien- cy, or vigor of the host. For some effect with large intensity, agreement could be obtained that the health effect was ad- verse. Between these two boundaries lies the domain of controversy. Efforts to rationalize the decision-making proc- ess and achieve consensus have used proba- bility estimation techniques (U.S. EPA, 1989~. Such efforts have the virtue of measuring the range of uncertainty and making the degree of agreement explicit. Table 28-1 lists various responses to lead, threshold blood lead concentrations for alterations (measures of sensitivi- ty), and estimates of specificity. The thresholds are based on a consensus of investigators doing research on responses to lead exposure. 1.0~ O , W\<~ / \ Ao Bo O CO Aloo Bloo ClooDo LEAD DOSE - Dtoo FIGURE 281 Distribution of frequencies of four putative responses to internal lead dose. A represents effect beg nning at very low doses; B. C, and D are effects at higher lead doses. D might represent death. No, Bo' Co and Do represent doses above which first response would occur. Moo, Moo, CIOO, and Deco represent doses above which all subjects show effects. Source: Needleman, 1987.

OCR for page 297
300 FIGI~E 28 2 Intensity of given effect as func- tion of dose. Thick part of line represents areas of disagreement as to whether health effect has occurred. Source: Needleman, 1987. MARKERS OF SUSCEPTIBILITY A given internal dose of lead does not produce an effect of the same intensity in all people. Young children are more vulnerable to effects, perhaps because of increased metabolic rate, differences in blood-brain barrier permeability, or stage of neurogenesis or synaptogenesis. Women appear more susceptible to effects of lead than men. Therefore, sex and age may be considered markers of vulnerability to lead. Rogan and colleagues (1986) studied black children and found that children with high free erythrocyte protoporphyrin had low delta-aminolevulinic acid dehy- drase and suggested that the enzyme system could be used as a marker. McIntire and Angle (1972) measured blood lead and glucose-6-phosphate dehydrogenase (G- 6-PD) activity in students living near smelters and found that G-6-PD-deficient subjects had significantly higher eryth- rocyte lead than those with normal G-6- PD content. That suggests that G-6-PD is a marker of susceptibility. G-6-PD is in- herited in a simple, Mendelian fashion. LL LL a an z z - ~as agreed / l~uPOn be 100% NEURODEVELOPMENTAL TOXICOLOGY Unsafe as agreed / upon by 100'/ ~ ' ~ 1 1 l 1 Lead dose resulting in earliest detectable etfec t METHODOLOGIC CONSIDERATIONS IN THE ESTABLISHMENT AND EVALUATION OF MARKERS OF DEVELOPMENT Differentiating Markers of Exposure and Effect At first glance, the difference between markers of exposure and markers of effect seems self-evident. However, the concen- tration of an internal dose of a toxicant, which seems clearly a marker of exposure, involves the organism's inherent respon- siveness to the stimulus (i.e., external dose of the agent). That becomes obvious when one considers that subjects exposed to the same external dose of a toxicant demonstrate different internal doses. A biochemical change that is highly cor- related with the dose of an agent can serve as a marker of exposure. For example, free erythrocyte protoporphyrin has been used in this regard by clinicians evaluat- ing lead exposure. An increase is a bio- chemical response of the heme system to altered activity of the enzyme ferrochela- tase; it represents an alteration in the functioning of an integrated biochemical system and shares the properties of a mark- er of effect. But, because it tracks lead dose predictably, it has been used as a marker of exposure. In addition to correlation with toxicant dose, two other attributes of a candidate exposure marker-sensitivity and specifi- city-influence its utility as an exposure

OCR for page 297
LE4D AS A PARADIGM TABLE 28 1 Responses to Lead 301 Threshold, Biologic Effect ug/dla Speahlatv Heme pathway effects S'Ps~ Decreased 6-am~nolevulinic and dehydrase acts Increased free e~ythrocyte protopo~phyr~n concentration Decreased heme production Basophilic stippling Frank anemia Central nervous system effects Altered electrophysiologic responses Quantitative BEG abnormalities appear Evoked potential abnormalities appear Psychologic deficits _ . . . . 10 10 15 40 50 60 35 10 Moderate Moderate Moderate Low High Low Low Low IQ deflate increased 15 Low Attention deflate appear 10 Low Speech and language deficits appear 35 Low Other effects 1,25-vitamin D hydroxylase activity altered Na-K ATP-ase activity altered 12 20 High Low Threshold blood lead concentration for indicating that an effect has occurred. marker. These concepts are discussed in detail in the appendix. Sensitivity is the probability that an effect will be detected by the marker. Specificity is the probability that the lack of effect will be correctly identified, i.e., that the marker does not falsely indicate an effect. Together, these two characteris- tics and the prevalence of the effect in the population determine the predictive value of the marker. The higher the specificity of a given marker (i.e., the lower its orobabilitv of association with other diseases or ex- posures), the greater its utility in diag- nosing disease or exposure. The specifici- ty of a marker of effect is related to its proximity in the causal chain to the health effect in question. That in turn depends on the definition of adverse health effect and on the process of causal inference. ADVERSE HEALTH EFFECTS Causal Inferences in Toxicant Exposure Establishment of a causal nexus between the exposure variable or marker and a dis- ease is essential in defining the relation- ships between the variables. In nonexperi- mental studies, investigators must rely on experimental design and model-building to determine the most reasonable causal model. Scientists must balance type I errors (accepting spurious causal hypo- theses as true) and type II errors (re- jecting true causal hypotheses as spur- ious). Avoiding type I errors has been emphasized; less attention has been given to the risk of type II errors. In making causal judgments, the following errors are often encountered: Overvaluing the importance of the criterion p = 0.05. Many investigators dis- miss findings in which the p is greater than 0.05. The use of this fixed boundary is not logical, but is a long-held prefer- ence that has become conventional. Improper causal specification or modeling. Many studies attempting to reduce con- founding control for many variables that might be markers of effect. That results in reducing the variance properly attri- butable to the causal agent under study.

OCR for page 297
302 Arguing for lack of effect from studies with inadequate power to find an effect. Power measures the probability that a study can find an effect if it is present. It is a function of three variables: the number of subjects studied, the effect size under scrutiny, and the significance criterion selected by the examiner. Many studies that have argued for a lack of causal effect have done so on the basis of sample sizes with less than a 50% chance of finding an effect at a statistical sig- nificance of 0.05. Assigningsharedvariancetotheconfound- er. If the main effect (e.g., exposure vari- able) is highly associated with another variable, the effect might not be detected NEURODEVELOPMENTAL TOXICOLOGY in regression analyses. This problem is referred to as collinearity of effect and confounders. The order in which the vari- ables are considered in the regression equation will affect the conclusions of the analyses. Evaluating studies in isolation. Many re- view articles discuss studies in isola- tion, tabulate their advantages and dis- advantages, and then count the studies that favor and do not favor a causal effect. That type of evaluation seriously degrades the data. A relatively new form of review, meta-analysis, or quantitative summaries of many studies, promises to improve on the evaluation of multiple studies.