Genetics and Nutrition
Studies have demonstrated remarkable genetic diversity among humans. No two individuals on this planet are alike genetically, except for identical twins, and even they vary because of somatic mutations in the immune system. The well-known individual uniqueness of physiognomic features extends to a variety of genetically determined biochemical and immunologic characteristics. Such traits include blood groups and tissue histocompatability antigen (HLA) types as well as enzymatic and other proteins. More recently, extensive variability in noncoding DNA has been described.
The variability of enzyme levels within the normal range in a population often has a simple genetic basisdifferent structural alleles at a gene locus that specify slightly different mean enzyme levels (Harris, 1975). The widely variable but unimodal distribution of activity for a given enzyme in a normal population may be the sum of overlapping curves, each characteristic of its underlying allele.
No general predictions regarding the impact of genetic variation on nutrition, health, and disease can be made. Every genetic system has a different evolutionary background and must be investigated separately.
Applicability to Nutrition
Assessments of human nutrition are not complete without consideration of the underlying genetic variability, which may be reflected as differences in nutritional processes such as absorption, metabolism, receptor action, and excretion (Velazquez and Bourges, 1984). Inborn differences in the activity of enzymes and other functional proteins contribute to variations in nutritional requirements and to the differential interaction of certain nutrients with genetically determined biochemical and metabolic factors. This inborn variation is quite different from epigenetic variation under conditions of growth, pregnancy, and old age. Genetic variation may also affect food likes and dislikes and, as a consequence, nutrition. For example, the inability to taste the synthetic chemical phenylthiocarbamide is a common monogenic trait that makes a large portion of the population unable to taste this chemical that others find quite bitter (Harris and Kalmus, 1941). Other examples, such as variability in the tasting of artificial sweeteners, are less well studied.
A key question of significance to nutrition is the extent of variation for a given gene product. Small variation can often be disregarded when making recommendations on nutrition, whereas wide variation cannot be ignored. The number of people affected by the variation is also important for policy setting. In the case of a monogenic variant, common genetic polymorphisms become important, especially when many people are affected. Rare variants affecting only a few people may pose
a problem if their health would be placed at risk under circumstances conferring benefits to the health of most of the population.
Nutritional factors may have played a role in human evolution by selecting for certain genotypes (Neel, 1984). Thus, periods of starvation may have favored genotypes predisposing to hyperlipidemia and non-insulin-dependent diabetes by allowing more ready mobilization of lipids and glucose that provided a slightly better chance of survival and reproduction. A similar reasoning applies to genotypes predisposing to obesity. Such a hypothesis would explain the relatively high frequency of these traits.
Inborn Errors of Metabolism as a Model
Even though inborn errors of metabolism are rare, their mechanisms may illustrate the role that more common genes may play in nutrition. The intrinsic processes required for proper nutrition, such as digestion, absorption, and excretion, are affected selectively by many different inborn errors of nutrient metabolism (Rosenberg, 1984). More than 200 such disorders have already been described. Among them are lactose intolerance (the inability to digest lactose); glucose-galactose malabsorption (the inability to absorb these nutrients); familial hypercholesterolemia, which develops in people who lack the receptors necessary to remove low-density lipoproteins (LDLs) from plasma; ornithine transcarbamylase deficiency, in which people lack an enzyme involved in the detoxification of ammonia; and hypophosphatemic rickets, in which the renal reabsorption of phosphorus and the intestinal absorption of phosphorus are impaired (Rosenberg, 1980). These disorders vary with regard to nutrients involved, frequency of occurrence, ethnic distribution, clinical severity, and disease manifestations (Holtzman et al., 1980). They may produce an internal or functional deficiency of an essential macro- or micronutrient despite adequate dietary intake; they may lead to chemical toxicity by blocking a catabolic pathway needed to metabolize an ingested nutrient; they may interfere with the formation of a needed product from an ingested nutrient; they may disrupt feedback regulatory pathways; or they may lead to pathological accumulation of macromolecules. Many of these metabolic disorders can be managed by modifying nutrient intake. For example, deficient intestinal absorption of a nutrient can be remedied by high oral intakes or parenteral administration; toxicity resulting from a blocked catabolic pathway of an essential amino acid can be relieved by restricting intake; vitamin supplementation may help to ameliorate a disturbance due to deficiency of an enzyme that requires the vitamin as a cofactor (Rosenberg, 1980).
Uptake of a variety of nutrients and other critical metabolites by cells is carried out by receptor-mediated endocytosis. The receptor that facilitates the uptake of LDLs (LDL receptor) has been studied in detail in the normal state and is an excellent model for receptor function in general (Goldstein and Brown, 1985). Some mutations of this receptor lead to defective transfer of LDLs into cells and increased LDL and cholesterol levels in the blood; this in turn predisposes to coronary heart disease (CHD). Mutations for the heterozygote state of familial hypercholesterolemia are found in approximately 1 in 500 people and predispose to premature coronary arteriosclerosis. Homozygotes are very rare (one in a million) and often develop CHD before 20 years of age (Goldstein and Brown, 1983).
These rare genetic disorders affecting enzymes and receptors illustrate how a severe genetic defect may lead to malnutrition or specific damage to a given organ system. They can serve as models for the study of milder but more common genetic variations in their effect on nutrition.
Possible Effects of Heterozygosity
The expression of most inborn errors of metabolism requires the presence of two identical mutant geneseach contributed by the carrier parent of an affected patient. The most common genetic disease of this sort is phenylketonuria, an autosomal recessive disease with a maximum frequency of 1 in 10,000 births. Most other inborn errors of metabolism have frequencies between 1 in 40,000 and 1 in 250,000 (Vogel and Motulsky, 1986). These rare inborn errors are not usually considered in making nutritional recommendations for the population as a whole, but carriers of the relevant mutant gene are quite common in the population. For example, 2% of the population are carriers of the mutant gene for phenylketonuria. In patients with inborn errors of metabolism, the involved enzyme has very little normal activity. Normal people have approximately 100% activity; carriers have about 50%. Under most conditions, a 50% level of enzyme activity is sufficient for adequate function. Thus, carriers are in good health. Under conditions of growth, stress, illness, or malnutri-
tion, however, even 50% of enzyme activity may not be sufficient, and specific abnormalities related to the underlying enzyme activity might result (Vogel and Motulsky, 1986). Familial hypercholesterolemia is a relatively common heterozygous condition (1 in 500), which in the homozygous state is very rare (1 in 1 million) (Goldstein and Brown, 1983). More research is required to define whether the carrier heterozygotes for rare inborn errors are at risk for disease.
Research Methods in Medical Genetics
The number of genes shared by family members depends on their degree of relatedness (Vogel and Motulsky, 1986). First-degree relatives (full siblings or parents and their children) share an average of 50% of their genes; second-degree relatives share an average of 25% of their genes. Unrelated members of an ethnic group or race with common ancestry share a common gene pool and, hence, may resemble one another (biochemically and physically) more than they resemble people from other groups.
Generally, proof that a trait is genetically determined comes initially from a demonstration of familial aggregation, i.e., the trait is more frequent among relatives of affected persons than it is in the general population. However, familial aggregation by itself does not prove the involvement of genetic factors. Environmental exposures and lifestyle are also shared by family members and may cause a higher frequency of a particular trait among relatives. Absence of correlation for a given trait among spouses living in the same household and among children and their adoptive parents or siblings (who share no genes) argues against common environmental factors when positive correlations for the trait are obtained for biologic relatives. Higher correlations among identical twins (who share all their genes) than among nonidentical twins (who share half their genes) also argue for genetic factors. However, because identical twins are likely to select similar environments, ideal studies attempt to assess identical twins reared apart. Since such twin pairs are not found frequently, study sizes are necessarily small (Vogel and Motulsky, 1986; see also Chapter 7).
Biometrically analyzed studies of identical twins reared separately may suggest the operation of genetic factors for a given trait, but they provide no information regarding the number of genes involved or the mechanisms of their action. Obesity is a good example. Studies of adoptive families and twins suggest that genetic factors are operative (see section on Obesity), but their mechanisms remain unknown.
The inheritance of single genes can be inferred from the nature of their segregation in families. Appropriate genetic-statistical techniques such as segregation analysis are available to test for the mode of inheritance. The principle of such methods is to test family data against various models of genetic transmission and find the best-fitting model (MacLean et al., 1985).
When the nature of the genetic abnormality is known for a monogenic or Mendelian trait, the abnormal or variant gene product can be studied in families, and appropriate segregation studies can be conducted. Where the specific structural defect in a protein or enzyme has been identified, the corresponding alteration in DNA can be inferred.
Much biologic insight into the role of genetics in humans has come from investigations of the role of specific genes and gene products. These approaches are reductionistic in nature, and their results ultimately need to be integrated with other observations of complex interaction of several genes. Furthermore, the role of special environmental factors in modifying gene action must be considered (see section on Gene-Environment Interaction, below).
Another approach to monogenic action is based on linkage analysis, the study of cosegregation of a common (or marker) gene with a physiologically or pathobiologically important gene in which the close physical apposition of these two genes on a given chromosome is studied. This type of study requires investigations of families in which an informative marker gene occurs in various related family members. Many DNA variants distributed over every one of the 23 chromosomes in humans are already known, and marker genes for closely spaced chromosomal sites are rapidly being elucidated. Because most DNA does not code for proteins, the variability of DNA is usually not expressed phenotypically, but DNA variants can be used as genetic markers to detect closely linked genes of physiological or pathological importance (Botstein et al., 1980).
Recently, some DNA mutations have been detected by linking harmless DNA variants to the diseased gene before any information about the nature of the defect was known. This approach has been referred to as reverse genetics (Orkin, 1986).
Population studies demonstrate different frequencies of genetic traits in various ethnic groups (Vogel and Motulsky, 1986). If the genetic trait is of physiologic or clinical significance, the total
impact of a given gene's variation may differ in different populations.
Genes do not act in a vacuum; the action of one gene may depend not only on other genes, but also on the environment. This principle is well illustrated by the field of pharmacogenetics. Certain inherited enzyme variants are harmless by themselves but may cause untoward effects in the presence of a drug that requires the normal variety of that enzyme for its inactivation. The presence of the enzyme variant without the drug and the administration of the drug to a person with the normal enzyme are harmless. If the drug is given to the carrier of the enzyme variant, however, a reaction to the drug ensues. Examples include hemolytic anemia from glucose-6-phosphate dehydrogenase deficiency, prolonged apnea from pseudocholinesterase variation, and various drug reactions associated with defective acetylation of drugs such as isoniazid (Stanbury et al., 1983).
The concept of pharmacogenetics has been widened to ecogenetics, i.e., the interaction of specific genetic traits with any environmental agent to produce a given effect. There are three ecogenetic examples with relevance to nutrition: (1) severe flushing of the skin after exposure to alcohol in many Orientals, who often lack an isozyme of aldehyde dehydrogenase, which is involved in the metabolism of alcohol (NIAAA, 1987); (2) hypertension in genetically predisposed persons who migrate from a primitive environment to a more westernized one (Page, 1979); and (3) gastrointestinal distress after moderate milk consumption by many people with genetically determined lactase insufficiency (see section on Lactose Malabsorption, below) (Lisker, 1984). There are many more examples. As more is learned about the nature and extent of genetic variability, its interaction with the environment, and its effect on disease resistance and susceptibility, it will become increasingly possible to advise genetically susceptible individuals about environmental factors (including diet) that will prevent various diseases determined by genetic-environmental interaction.
Other Genetic Factors that Affect Nutrition
In the central nervous system, genetic variation probably affects perception of taste, degree of satiation, and other factors likely to affect food intake. However; no critical data on humans exist in this area. Absorption can also be affected. Examples include increased iron absorption in hemochromatosis (see section on Hemochromatosis, below) and genetically determined absence of gastric intrinsic factor, which leads to defective vitamin B12 absorption and pernicious anemia (Velazquez and Bourges, 1984).
Ethnic and racial factors also require consideration. Relatives share common ancestors and therefore are more likely to share similar genes derived from that ancestor. In a sense, an ethnic group is an extended family, and similar considerations apply. Often, therefore, frequencies of genetically determined traits or diseases will differ among races (e.g., Caucasians, blacks, or Orientals), and even among ethnic groups of the same race. It may not be apparent whether an ethnic or racial difference for a given trait or disease is caused by the existence of different genes or because the unequally affected racial group lives in a different environment. However, if the presence of a gene or genes can be demonstrated, the differentiation between genetic and environmental factors usually becomes clear. Thus, although we strongly suspect that genetic factors cause the difference between Pima Indians and Caucasians in the frequency of obesity, we cannot be absolutely sure since we have no gene marker. On the other hand, there is little question that the difference in frequency of hypolactasia in blacks and whites has a genetic cause, since tests for hypolactasia exist.
Such ethnic or racial differences may have policy implications; a nutritional policy may be desirable for one population group but would cause health problems in another.
Genetic Factors in Some Chronic Diseases
Coronary Heart Disease and Lipids
Hypercholesterolemia, high LDL levels, low levels of high-density lipoprotein (HDL), and low apolipoprotein A1 levels have all been implicated as risk factors for CHD, and all are influenced by both genetic and environmental factors. The presence of two different but common alleles at the apolipoprotein E locus has a large effect on cholesterol and apolipoprotein B levels (Sing and Davignon, 1985). The transmission of familial hypercholesterolemia due to various LDL receptor defects is Mendelian (Goldstein and Brown, 1983), and genetic factors seem to play a role in most
hyperlipidemias. A common condition known as familial combined hyperlipidemia appears to be transmitted as a monogenic autosomal dominant trait and is usually associated with elevation of apolipoprotein B (Brunzell and Motulsky, 1984). The etiology of this condition appears to be heterogenous. Various DNA markers of the apolipoprotein loci have been associated with hyperlipidemia, or CHD, or both, but findings are not entirely consistent (Deeb et al., 1986). Considerable research is under way in these areas and is likely to help in clarifying the exact contribution of genetic factors to the hyperlipidemias.
A crucial issue is the relevance of such data in the development of nutritional advice for the population. Some argue that regardless of genetic variation, a lowering of the cholesterol level of Western populations by dietary modification would substantially reduce the frequency of CHD. Those with the lowest blood cholesterol levels have the lowest mortality from CHD, which increases progressively with higher cholesterol levels (Martin et al., 1986). Most CHD deaths and most excess coronary events related to elevated cholesterol levels do not occur at the upper end but in the center of the distribution curve (see Chapter 7). What will be the effects of lowering cholesterol levels in the population? For example, reducing an individual's cholesterol level from 226 to 210 mg/ dl (which could be achieved through dietary modification) decreases the absolute risk of coronary mortality only slightly. However, a small reduction in absolute risk for an individual may translate into a major effect in the large number of people who constitute the population. In a hypothetical example, one can assume that by dietary modification, a person can reduce his or her risk of having a myocardial infarct within a given time span from 1 in 80 to 1 in 100. Applied to 100,000 persons, the expected frequency of 1,250 myocardial infarcts (1/80) would be reduced to 1,000 heart attacks (1/ 100). This 20% reduction (or prevention of 250 heart attacks) would be of significant public health importance.
Little is known about the effect of nutrition-genetic interactions on lipids. The role of the LDL receptor, structural variation in apolipoproteins B and E, the regulation of hepatic apolipoprotein B synthesis, and many other factors need to be studied to resolve these issues. Further knowledge in this important area is required to assess the role of diet in reducing lipid levels in individuals. It is certain, however, that genetic factors play an important role in determining cholesterol levels.
Response to dietary restriction of saturated fats and cholesterol or lipid levels will be variable, and not everyone will benefit equally from dietary moderation. Some people will be sensitive, whereas lipid levels in others may be resistant to control by dietary changes.
Hypertension and Salt
Blood pressure levels are under strong genetic control, as shown by studies in families, adopted children, and twins (Burke and Motulsky, 1985). There is good evidence for the role of sodium in the causation and maintenance of high blood pressure. In populations, the frequency of high blood pressure is related to average sodium intake. However, sodium loading does not cause elevation of blood pressure in all people, and sodium restriction lowers blood pressure in many but not all hypertensives. Hypertension is an ecogenetic trait in that so-called primitive populations have little hypertension. High blood pressure develops in some people when they translocate to a Western-type environment that includes a high salt intake. Populations of African origin in the United States have a higher mean blood pressure and a higher frequency of hypertension than those of European origin. Recently discovered differences between black and white hypertensive populations include the absence of elevated red-cell sodium/lithium countertransport in blacksa finding that is common among white hypertensives. This transport trait appears to be under monogenic control (Motulsky, 1987a). This and other evidence suggests that hypertension is a heterogeneous entity with different genetic mechanisms.
Salt restriction has been advocated to reduce the frequency of high blood pressure but may not be equally helpful for all individuals (see Chapters 15, 20, and 28). In addition, clinical or laboratory criteria needed to characterize the degree of salt-sensitivity among people are lacking.
Most cases of diabetes mellitus in humans fall into one of two categories: noninsulin-dependent diabetes mellitus (NIDDM) and insulin-dependent diabetes mellitus (IDDM). Twin and family studies strongly indicate the existence of genetic factors in both varieties, but the exact mechanisms remain unknown. More is known about some of the specific genes involved in IDDM (e.g., HLA-related genes) than in the more common NIDDM
variety. The very high concordance of identical twins for NIDDM suggests that genes play a more central role in this disease (Barrett et al., 1981). Although nutrition and dietary components are important in the treatment of clinical diabetes, they do not appear to be involved in the pathogenesis of either form of the disease (Glinsmann et al., 1986), aside from the clear relationship of NIDDM to total caloric intake.
The development of obesity is not simply a matter of caloric intake but involves genetic factors as well. This has been shown in studies of humans and of animals.
It has been known for many decades that obesity is a familial trait. Most obese patients have at least one obese parent; however, members of a family or relatives share many common aspects of nutrition and environment. For example, in the analysis of the Ten State Nutrition Survey (Garn and Clark, 1976), other factors such as socioeconomic class, age, and sex were prominent predictors of obesity (Garn and Clark, 1976). More recently, Garn (1985) noted that most children of obese parents become overweight, whereas those of lean parents are thinner.
The most compelling evidence for genetic factors comes from studies of monozygotic and dizygotic twins (Borjeson, 1976; Bouchard et al., 1985; Fabsitz et al., 1980; Feinleib et al., 1977; Medlund et al., 1977; Stunkard et al., 1986a,b). Borjeson (1976) studied 40 monozygotic twin pairs and 61 same-sex dizygotic twin pairs and estimated heritability of obesity to be 88%. (Heritability may range from 0%, i.e., no genetic factors, to 100%, indicating that a trait is determined entirely by genetics). A similar conclusion was reached by Brook et al. (1975). A much larger twin study based on the large Swedish Twin Registry showed a high concordance of body fatness in monozygotic but not in dizygotic twins (Fabsitz et al., 1980). A potent interaction between genetic predisposition and environment is suggested by a recent feeding and exercise challenge study in monozygotic and dizygotic twins (Poehlman et al., 1986).
The relative roles of genetics and the environment have also been examined by comparing obesity among adoptees to obesity in their adoptive and biologic parents. In several early adoption studies, no clear trends were found (Annest et al., 1983; Withers, 1964). In at least one study, obesity was found among children with obese adoptive parents (Annest et al., 1983). In most of these studies, no information on biologic parents was available. Children may have spent an extended period with a biologic parent before adoption, and adoptive placement may have been based on physical similarity between adoptive parents and the adopted childa common practice several decades ago. However, two recent studies, one based on a Danish adoption registry (Stunkard et al., 1986a) and another on fatness data from Iowa (Price et al., 1987), included information on biologic parents as well as adoptive parents. These show a stronger correlation in body mass index (BMI) between adoptees and their biologic parents than between adoptees and their adoptive parents. In the Iowa study, the correlation of BMI was independent of height and applied to the full range of obesity. In stepwise multiple regression analyses, the best correlations were predicted by a model in which genetic and nonfamily environmental factors were both considered. Even so, much of the variance was unexplained. Additional studies examining biologic parents and siblings have shown a variable genetic and genetic-environmental contribution (Bouchard, 1988). Nevertheless, taken together, the results of such studies suggest that multifactorial polygenic factors play a role in human obesity. This formulation does not rule out the involvement of major genes that have yet to be identified.
Ethnic differences also are consistent with, but do not prove, the role of genes in obesity. The higher frequency of overweight among black women compared to white women may at least partially be caused by genetic factors (Van Itallie, 1985). This may also apply to Mexican-American children, who are fatter than either white or black children (Mueller, 1988). The high prevalence of obesity among several Native American tribes and Pacific Island populations is difficult to explain by excess calories alone.
There is evidence from a study of twins (Bouchard, 1988) and from several studies of ethnic groups (Mueller, 1988) that patterns of body fat distribution are inherited. This is especially apparent among children. For example, Mexican-American children have a pronounced upper body fat distribution, which is sometimes independent of adiposity per se (Mueller, 1988), whereas children of European origin seem to have a more peripheral distribution of fat (Mueller, 1988).
The nature of the specific genes causing such phenomena is still unknown.
Obesity in mice and rats can be caused by several different single-gene mutations (Bray and York, 1971). Some strains of pigs and rats become obese under some feeding circumstances in which control animals do not. The exact biochemical' or metabolic defects responsible for this have not yet been discovered. Intensive research is under way to clone relevant candidate genes.
It is unlikely that any of the animal models of monogenic obesity will fully explain the genetic factors operative in human obesity, since the genetic mechanism of human obesity does not appear to be monogenic. It is likely, however, that some of the genes involved in monogenic animal obesity may play some role in human obesity. Information on the nature of the genes involved in animal obesity will allow direct testing of the involvement of similar genes in humans by using techniques of molecular genetics.
Certain types of cancer appear to be inherited. It is becoming increasingly apparent, however, that most cancers can be attributed to interactions between genetic (hereditary, or endogenous) factors and environmental (exogenous) factors.
It is helpful to consider causes of cancers in three broad groups: genetic, genetic-environmental, and environmental alone. The first group consists of cancers that appear to be determined largely by genetic factors with little or no environmental influence. One example is the high risk for colon cancer in people with familial polyposis of the colon. Other cancers in this group appear to be associated with impaired DNA repair mechanisms, such as in xeroderma pigmentosum and Bloom's syndrome. These examples illustrate a merging with the second group; for example, the development of malignancies in people with xeroderma pigmentosum requires exposure to an environmental agent (ultraviolet light) in predisposed individuals.
For the second group of cancers, genetic and environmental factors appear to be importantindependently or synergistically. Examples include cancers that tend to run in families, such as some cancers of the stomach, colon, and breast. There is evidence that in colon cancer and possibly other cancers in this category, the pathogenic mechanism may involve the genetic transmission of a recessive gene that is present in all body cells and by itself does not cause cancer (Bodmer et al., 1987). A somatic mutational event of the allelic partner in a single colon cell causes homozygosity at this locus and frees this cell of growth restraint, ultimately causing clinical colon cancer. Different environmental agents, including metabolites derived from food, could cause the second somatic mutation (Knudson, 1985; Solomon et al., 1987).
Cancers in the third group are produced largely by environmental agents, which in general are independent of genetic variation. This category includes most malignant neoplasms. For some of these, however, genetic variation may affect cancer risks, for example, by altering the metabolism of carcinogens. Such mechanisms may affect individual risks of some of the smoking-related cancers (Ayesh et al., 1984; Mulvihill, 1976).
No studies of cancer risk in humans have yet indicated a relationship between dietary factors and genetic factors; however, such a relationship may be responsible for the increased cancer risk in genetically predisposed individuals. Therefore, there is some reason to postulate that dietary changes could have at least as much effect on cancers dependent on a genetic mechanism as they have on cancers that are free of genetic influence.
At the time of birth, all humans (and all other mammals) produce the intestinal enzyme lactase to metabolize lactosethe main constituent of milkinto glucose and galactose. In most humans, the ability to digest lactose disappears after weaning due to progressive decline in intestinal lactase activity; such individuals are often referred to as lactose malabsorbers. Undigested lactose in the gastrointestinal tract of such people is decomposed by bacteria, causing bloating, diarrhea, intestinal rushes, flatulence, and even nausea and vomiting in severe cases. However, some peopleparticularly those of European origindo not lose this ability and have persistent intestinal lactase activity (Lisker, 1984). This persistence of lactose absorption is controlled by a gene for persistence of intestinal lactase activity (L). People who do not carry this gene, and therefore cannot digest lactose after weaning, are homozygotes (11) at this locusthe usual status for the majority of the world population. Milk drinking does not induce intestinal lactase activity in thosewho no longer have this capacity, nor will lactose restriction reduce the intestinal lactase activity among those who never lost it. Lactose absorption or malabsorption is an inborn genetic trait. Acute or chronic gastrointes-
tinal disease may cause secondary hypolactasia among people with persistent lactase activity, but such activity returns after the illness. Most populations throughout the world have hypolactasia of the genetic variety. Only populations from central and northwestern Europe and from areas in Africa with a long history of dairy farming have high frequencies of persistent lactase activity. Presumably, the gene for lactase persistence had a survival advantage in dairy farming cultures and over the generations increased in frequency because people who were able to absorb milk as children and young adults were either more fertile or less likely to die early.
Milk is not consumed widely in populations where lactose malabsorption is common. Since milk is an important source of protein, calcium, and riboflavin, the existence of the lactase polymorphism has policy implications for the use of milk as a supplement for the world's lactose malabsorbers. However, most lactose malabsorbers can drink at least 250 ml of milk without much difficulty. Further studies are required to define more fully the amounts of lactose tolerated by malabsorbers.
The clinical manifestations of primary hemochromatosis (including liver and heart damage, arthritis, diabetes, and skin discoloration) are caused by the toxic effects of excessive iron stores in many organs following increased iron absorption over many years. Nonspecific symptoms such as weakness and fatigue are frequent.
Affected persons are homozygotes for a gene that facilitates increased iron absorption and is carried on the short arm of chromosome 6, closely linked to HLA Locus A (Bothwell et al., 1983). The homozygous state affects from 1 in 600 to 1 in 1,000 people in the United States and western Europe, indicating that approximately 7% of the population are heterozygotes or gene carriers for the condition. Clinically apparent disease is more common among males, since females can eliminate some excess iron in their menses. The increasing practice of measuring serum iron and iron saturation is leading to the detection of cases who are not yet affected clinically but who will develop clinically apparent disease at a later stage. The extent of increased iron storage in hemochromatosis can be estimated by serum ferritin measurements.
Since iron deficiency is common among the general population, additional supplementation of flour with iron has been recommended by public health authorities and is practiced in Sweden. The onset of clinical hemochromatosis in homozygotes presumably would be hastened by such a process. However, since the homozygous hemochromatosis genotype is found in, at most, 1 in 500 people, some observers believe that iron supplementation benefits a much larger fraction of the population (primarily women and children) and outweighs the damaging effects of iron for homozygotes. A crucial issue in this connection is related to the. iron absorption status of the very common hemochromatosis heterozygotes (i.e., close to 1 in 10 persons). However, although liver iron stores increase somewhat among male heterozygotes as they get older, there is no evidence that heterozygotes are at risk for clinically apparent iron toxicity. Furthermore, there is currently no test for detecting heterozygotes in population studies.
The benefits of iron supplementation should be considered in the context of possible risk to occasional hemochromatosis homozygotes, especially since the health effects of mild to moderate iron deficiency are not as fully defined as desirable (Cook and Lynch, 1986).
It has been known for many years that alcoholism is a familial occurrence, but the role of genetic and environmental factors within the family have been difficult to separate. Evidence accumulated over the past two decades suggests that alcoholism results from the interaction of heredity and the environment (NIAAA, 1987; Motulsky, 1987a; Omenn, 1987). Data on the genetic contribution comes from studies of familial alcoholism (Cloninger et al., 1978; Winokur et al., 1970), twins (Kaij, 1960; Loehlim, 1972), adoptees separated from their biologic parents at an early age (Bohman, 1978; Goodwin 1987), and animal breeding studies (Crabbe et al., 1981; Thurman, 1980).
The phenotype of ''alcoholism" is heterogeneous. Alcohol-seeking behavior must be differentiated from alcohol dependence and alcohol tolerance. Target organ damage, such as that seen in Korsakoff's syndrome or alcoholic cirrhosis, poses genetic problems that are different from alcoholism (see Chapter 16).
Several studies of adopted children indicate that natural sons of alcoholics are three to four times more likely to be alcoholic than are natural sons of nonalcoholics, regardless of whether the sons were raised by their alcoholic biologic parents or by nonalcoholic adoptive
parents (Goodwin, 1987). Studies of adoptees in Sweden identify two types of genetic disposition to alcoholism: milieu limited and male limited. Milieu-limited susceptibility occurs in both sexes and needs environmental challenge for its expression. The alcoholism is usually not severe, has a late onset, and is often associated with minor law violations. Male-limited susceptibility occurs in the biologic fathers of adoptees, is highly heritable, is reflected in severe, early-onset alcoholism that often requires extensive treatment, and is often associated with major law violations.
Studies of twins have not provided as clear evidence for genetic factors as have the adoption studies. Several studies have shown higher concordance in identical twins than in nonindentical twins, whereas others have not (NIAAA, 1987).
Efforts are under way to identify neurophysiological, neuropsychological, and biochemical markers of genetic susceptibility to alcoholism. Characteristic electrical brain patterns have been found in nonalcoholic offspring of alcoholic fathers (Mendelson and Mello, 1979). Fast electroencephalographic (EEG) activity and deficiencies in alpha, theta, and delta EEG activity have been reported in sons of alcoholics (Gabrielli et al., 1982). Studies of event-related potentials in the nonalcoholic sons of alcoholic fathers showed that they had a decreased amplitude in their P3 wave similar to that found in abstinent alcoholics (Begleiter et al., 1980). Evidence suggests that decreased P3 amplitude precedes the onset of alcohol abuse and may be a genetic marker of susceptibility (NIAAA, 1987). Tests of abstracting, problem solving, perceptual-motor functioning, and stimulus augmenting showed that nonalcoholic men with a family history of alcoholism performed less well than controls with no such family history (Schaeffer et al., 1984).
Research on biochemical markers has included the study of genetic variation in alcohol-metabolizing enzymes, especially alcohol dehydrogenase and aldehyde dehydrogenase. These enzymes help to eliminate alcohol from the body, mainly by oxidative metabolism in the liver. However, markers for these enzymes have not been associated with predisposition to alcoholism (NIAAA, 1987).
Oriental populations have a much higher frequency of the aldehyde dehydrogenase polymorphism. Carriers of this genetic trait flush more readily on exposure to alcohol. It is likely that this flushing reaction is a deterrent to excess alcohol consumption (NIAAA, 1987).
Policy Implications of Genetic Variations
Public health policies generally focus on the average human being, without considering genetic variation. Thus, dietary recommendations usually provide a sufficient nutrient intake even for those with the highest requirements. This policy appears sound when variation for a given nutrient is small. If special dietary requirements affect only very few people with an inborn error of metabolism, policy recommendations could legitimately ignore such outliers, since such people can be identified and treated by physicians. In many instances, however, the true extent of genetic variation is unknown. The frequency of genetic variants in enzymes and proteins and their effect on enzyme activity suggest that there may be great variations in nutritional requirements or in gene-nutrition interaction.
Genetic factors as they relate to nutrition must be considered individually. The role of genetics in relation to dietary lipids and CHD is one example. High-fat intake constituting 40% of the calories in many Western diets and with a high saturated fat content leads to relatively high lipid levels and high frequencies of CHD. Serum cholesterol at all levels has been correlated with CHD frequency, but approximately 40% of all coronary events occur in the population with the highest 25% of cholesterol levels, including many genetic hyperlipidemias (see Chapter 7).
Many coronary events could be prevented if the entire population would reduce its cholesterol levels by decreasing saturated fat intake. Although individual risk reduction for those with "normal" cholesterol levels is relatively small, the total effect on the population would be considerable, because even small reductions in risk lead to a large public health benefit when the entire population is considered. This phenomenon of small effects in individuals, but a large impact on the population, has been termed the prevention paradox. A double-pronged strategy of case detection and individual management of those at high risk, together with a public health strategy to reduce cholesterol levels in the population as a whole, could therefore have large effects.
Currently, the adverse public health effects of obesity are well known. Thus it is appropriate to advise everyone to avoid obesity. In the future, it may be possible to identify those at particular risk of obesity from excess caloric intake and to focus special preventive efforts on this group (i.e., help them to avoid gaining weight).
For many diseases, there are substantial racial or ethnic differences in frequencies. Thus, a recommendation for a population subgroup may need to differ from that for the racial majority, though this may raise difficult policy questions because the recommendation can be easily misunderstood. However, there are precedent examples unrelated to nutrition including greater screening for Tay-Sachs disease among Ashkenazi (Eastern European) Jews, thalassemia among Mediterranean and Southeast Asian populations, and sickle-cell disease among blacks.
Problems can arise when a small group of persons is placed at high risk by a policy decision that would benefit the majority. An especially poignant example relates to iron supplementation and its effect on people with hemochromatosis (see section on Hemochromatosis).
Public health policy designed to benefit the population without discernible risk to certain individuals or subgroups has many merit it is simple to implement, may cost less to society, and may benefit many people. Examples of such policies are vaccination against poliomyelitis even though some are genetically susceptible to paralysis (Motulsky, 1987b) or fluoridation of the entire water supply to prevent caries even though there are genetic differences in caries susceptibility. As knowledge of genetic variation and its impact grows, it may be possible to direct recommendations to individuals, and it may no longer be necessary to assume the existence of an average person who can respond to general recommendations. A more sophisticated form of disease prevention based on biologic variation will then be desirable.
Genetic variability in biochemical processes is ubiquitous; every person is genetically unique. The relevance of this genetic individuality for nutrition and for the role of certain nutrients in disease causation requires much greater understanding. Processes of absorption, enzyme digestion, biosynthesis, catabolism, transport across cell membranes, uptake by cell receptors, storage, and excretion all vary extensively, and the variation is often known to be genetically determined, at least in part. The importance of such variation for nutrition needs to be determined by further studies in families Inborn errors of metabolism affecting these processes are genetic diseases that cause nutritional disorders of various kinds. Their rarity makes them medical problems; they are not usually viewed as problems of nutrition that affect public health. Heterozygotes, or carriers, of these inborn errors are much more common in the population. The possible role of the carrier state in causing clinically manifest disease during periods of stress, infection and malnutrition, for example, requires more study.
Most chronic diseases whose etiology and pathogenesis are influenced by nutritional factors have genetic determinants. High blood pressure, obesity, hyperlipidemia, atherosclerosis, and various cancers appear to aggregate in families for genetic reasons rather than merely because of a common environment. Recommendations to avoid nutrient excesses that predispose to these diseases are therefore unlikely to apply to everyone in the same way, and poorly understood interactions between genetics and the environment often govern the outcome of suboptimal nutrition. For most diseases, we lack the knowledge needed to identify susceptible genotypes by appropriate tests. For other conditions, however, such as the hyperlipidemias, we can already identify persons at high risk and concentrate specific medical efforts on this subpopulation. For conditions where specific tests are lacking but there is a strong family history of a given disease, appropriate preventive approaches can be tried for family members who have not yet been affected.
With advances in knowledge, an increasing number of population subgroups will be found to be at higher, or lower, risk for one or another chronic disease because of their genetic makeup. Specific recommendations directed at high-risk populations therefore will become possible and desirable. In the meantime, dietary recommendations directed at the entire population are appropriate for many conditions, even though different people will benefit unequally from such advice. As an example, a diet low in both fat and salt is likely to reduce disease incidence in the general population even though the beneficial effects for a given person may be small or nil. This paradox explains some of the controversies between those who promote an approach based on attention to persons at high risk and those who support a more global public health approach. Both approaches are needed. The high-risk approach is most appropriate in medical practice. The population approach requires support by the media, the food industry, nutritionists and dietitians, the public health profession, and the medical profession. When using the population approach, however, one must guard against harming certain genetically variant persons.
As we acquire a better understanding of genetics and it becomes increasingly possible to investigate genetic phenomena at a fundamental level, we will be able to conduct better studies to elucidate the exact role of genetic factors for each nutrient and
for various diseases. As these roles become understood, the associations between genetic factors and the interactions between heredity and the environment will become clearer and may lead to the modification of current policies.
Directions for Research
· Investigate the extent of genetic variability in requirements for a given nutrient and whether there are racial or ethnic differences.
· Determine whether the heterozygous state for certain inborn errors of metabolism is a marker for increased risk for disease.
· Define the interaction between heredity and the environment for nutrients, e.g., whether persons with certain common genotypes react in unusal ways to certain foods.
· Define salt sensitivity in hypertensives and simple methods for its detection.
· Learn how various ions (e.g., those of sodium, potassium, and chlorine) affect blood pressure and the role of genetic variation.
· Study why blacks have a higher frequency of high blood pressure, i.e., the role of genetic-environmental variation.
· Elucidate the role of genetics in dietary responsiveness to lipids.
· Define genes that affect parameters of lipid metabolism (e.g., HDL, LDL) and their interactions among themselves and with various nutrients.
· Conduct family studies to learn about mechanisms of racial differences in osteoporosis.
· Determine the basic genetic defect responsible for hemochromatosis.
· Study genetics of taste and olfaction.
· Delineate specific genes involved in NIDDM.
· Delineate genes involved in alcoholism.
· Define genes involved in human obesity.
· Conduct research on genetic variability and responsiveness to given dietary nutrients.
· Conduct research on the molecular basis of specific disorders responsive to diet. For example, the basis of salt sensitivity in hypertension, hyperlipidemia, osteoporosis, hemochromatosis, NIDDM, alcoholism, and obesity are currently amenable to further study.
Annest, J.L., C.F. Sing, P. Biron, and J.G. Mongeau. 1983. Family aggregation of blood pressure and weight in adoptive families. III. Analysis of the role of shared genes and shared household environment in explaining family resemblances for height, weight, and selected height/weight indices. Am. J. Epidemiol. 117:492-506.
Ayesh, R., J.R. Idle, J.C. Richie, M.J. Chrothers, and M.R. Hetzel. 1984. Metabolic oxidation phenotypes as markers for susceptibility to lung cancer. Nature 312:169-170.
Barrett, A.H., C. Eff, R.D.G. Leslie, and D.A. Pyke. 1981. Diabetes in identical twins: a study of 200 pairs. Diabetologia 20:87-93.
Begleiter, H., B. Porjesz, and M. Tenner. 1980. Neuroradiological and neurophysiological evidence of brain deficits in chronic alcoholics. Acta Psychiatr. Scand. 286:3-13.
Bodmer, W.F., C.J. Bailey, J. Bodmer, H.J.R. Bussey,. A. Ellis, P. Gorman, F.C. Lucibello, V.A. Murday, S.H. Rider, P. Scambler, D. Sheer, E. Solomon, and N.K. Spurr. 1987. Localization of the gene for familial adenomatous polyposis on chromosomes. Nature 328:614-616.
Bohman, M. 1978. Some genetic aspects of alcoholism and criminality. A population of adoptees. Arch. Gen. Psychiatry 35:269-276.
Borjeson, M. 1976. The aetiology of obesity in children. Acta Paediatr. Scand. 65:279-287.
Bothwell, T.H., R.W. Charton, and A.G. Motulsky. 1983. Idiopathic hemochromatosis. Pp. 1269-1298 in J.B. Stanbury, J.B. Wyngaarden, D.S. Fredrickson, J.L Goldstein, and M.S. Brown, eds. The Metabolic Basis of Inherited Disease, 5th ed. McGraw-Hill, New York.
Botstein, D., R.L. White, M. Skolnick, and R.W. Davis. 1980. Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am. J. Hum. Genet. 32:314-331.
Bouchard, C. 1988. Inheritance of human fat distribution. Pp. 103-125 in C. Bouchard and F.E. Johnson, eds. Fat Distribution During Growth and Later Health Outcomes. Alan R. Liss, New York.
Bouchard, C., R. Savad, J.-P. Depus, A. Tremblay, and C. Leblanc. 1985. Body composition in adopted and biological siblings. Hum. Biol. 57:61-75.
Bray, G., and D.A. York. 1971. Genetically transmitted obesity in rodents. Physiol. Rev. 51:598-646.
Brook, C.G.D., R.M.C. Huntley, and J. Slack. 1975. Influence of heredity and environment in determination of skinfold thickness in children. Br. Med. J. 2:719-721.
Brunzell, J.D., and A.G. Motulsky. 1984. Status of "familial combined hyperlipidemia." Pp. 541-548 in D.C. Rao, R.C. Elston, L.H. Kuller, M. Feinleib, C. Carter, and R. Havlik, eds. Genetic Epidemiology of Coronary Heart Disease: Past, Present, and Future. Alan R. Liss, New York.
Burke, W., and A.G. Motulsky. 1985. Hypertensionsome unanswered questions. J. Am. Med. Assoc. 253:2260-2261.
Cloninger, C.R, K.O. Christiansen, T. Reich, and I.I. Gottesman. 1978. Implication of sex differences in the prevalences of antisocial personality, alcoholism, and criminality for familial transmission. Arch. Gen. Psychiatry 35:941-951.
Cook, J.D., and S.R. Lynch. 1986. Review: the liabilities of iron deficiencies. Blood 68:803-809.
Crabbe, J.C., D.K. Gray, E.R. Young, J.S. Janowsky, and H. Rigter. 1981. Initial sensitivity and tolerance to ethanol in mice: correlations among open field activity, hypothermia, and loss of righting reflex. Behav. Neurol. Biol. 33:188-203.
Deeb, S., A. Failor, B.G. Brown, J.C. Brunzell, J.J. Albers, and A.G. Motulsky. 1986. Molecular genetics of apolipoproteins and coronary heart disease. Pp. 403-409 in Cold
Spring Harbor Symposia on Quantitative Biology, Vol. LI. Molecular Biology of Homo Sapiens. Cold Spring Harbor, New York.
Fabsitz, R., M. Feinleib, and Z. Hrubec. 1980. Weight changes in adult twins. Acta Genet. Med. Gemellol. 29:273-279.
Feinleib, M., R.J. Garrison, R. Fabsitz, J.C. Christian, Z. Hrubec, N.O. Borhani, W.B. Kannel, R. Rosenman, J.T. Schwartz, and J.O. Wagner. 1977. The NHLBI twin study of cardiovascular disease risk factors: methodology and summary of results. Am. J. Epidemiol. 106:284-295.
Gabrielli, W.F., Jr., S.A. Mednic, J. Volavka, V.E. Pollock, F. Schulsinger, and T.M. Itil. 1982. Electroencephalograms in children of alcoholic fathers. Psychophysiology 19:404-407.
Garn, S.M. 1985. Continuities and changes in fatness from infancy through adulthood. Curr. Prob. Pediatr. 15:1-47.
Garn, S.M., and D.C. Clark. 1976. Trends in fatness and the origins of obesity: Ad Hoc Committee to Review the Ten-State Nutrition Survey. Pediatrics 57:443-456.
Glinsmann, W.H., H. Irausquin, and Y.K. Park. 1986. Evaluation of health aspects of sugars contained in carbohydrate sweeteners: report of Sugars Task Force, 1986. J. Nutr. 116:SI-S216.
Goldstein, J.L., and M.S. Brown. 1983. Pp. 672-712 in J.B. Stanbury, J.B. Wyngaarden, D.S. Fredrickson, J.L. Goldstein, and M.S. Brown, eds. The Metabolic Basis of Inherited Disease, 5th ed. McGraw-Hill, New York.
Goldstein, J.L., and M.S. Brown. 1985. The LDL receptor and the regulation of cellular cholesterol metabolism. J. Cell Sci. 3:131-137.
Goodwin, D.M. 1987. Adoption studies of alcoholism. Pp. 3-20 in H.W. Goedde and D.P. Agarwal, eds. Genetics and Alcoholism. Alan R. Liss, New York.
Harris, H. 1975. The Principles of Human Biochemical Genetics, 2nd ed. North-Holland, Amsterdam. 473 pp.
Harris, H., and H. Kalmus. 1941. The measurement of taste sensitivity to phenylthiocarbamide. Ann. Eugen. 15:24-31.
Holtzman, N.A., M.L. Batshaw, and D.L Valle. 1980. Genetic aspects of human nutrition. Pp. 1193-1219 in R.S. Goodhard and M.E. Shils, eds. Modem Nutrition in Health and Disease, 6th ed. Lea & Febiger, Philadelphia.
Kaij, L. 1960. Alcoholism in Twins: Studies on the Etiology and Sequels of Abuse of Alcohol. Almqvist and Wiksell, Stockholm. 144 pp.
Knudson, A.G. 1985. Hereditary cancer, oncogenes and antioncogenes. Cancer Res. 45:1437-1443.
Lisker, R. 1984. Lactase Deficiency. Pp. 93-104 in A. Velazquez and H. Bourges, eds. Genetic Factors in Nutrition. Academic Press, Orlando, Fla.
Lochlim, J.C. 1972. An analysis of alcohol-related questionnaire items from the National Merit Twin Study. Ann. N.Y. Acad. Sci. 197:117-120.
MacLean, C.J., N.E. Morton, and R. Lew. 1985. Efficiency of lod scores for representing multiple locus linkage data. Genet. Epidemiol. 2:145-154.
Martin, M.J., S.B. Hulley, W.S. Browner, LH. Kuller, and D. Wentworth. 1986. Serum cholesterol, blood pressure, and mortality: implications for a cohort of 361,662 men. Lancet 2:934-936.
Medlund, P., R. Cederlöf, B. Floderus-Myrhed, L. Friberg, and S. Sörensen. 1977. A new Swedish twin registry containing environmental and medical baseline data from about 14000 same sexed pairs born 1926-1958. Acta Med. Scand. Suppl. (600):1-111.
Mendelson, J.H., and N.K. Mello. 1979. Biologic concomitants of alcoholism. N. Engl. J. Med. 301:912-921.
Motulsky, A.G. 1987a. Genetics and environment. Pp. 327-329 in H.W. Goedde and D.P. Agarwal, eds. Genetics and Alcoholism. Alan R. Liss, New York.
Motulsky, A.G. 1987b. Human genetic variation and nutrition. Am. J. Clin. Nutr. 45:1108-1113.
Mueller, W.H. 1988. Ethnic differences in fat distribution during growth. Pp. 127-145 in C. Bouchard and F.E. Johnson, eds. Fat Distribution During Growth and Later Health Outcomes. Alan R. Liss, New York.
Mulvihill, J.J. 1976. Host factors in human lung tumors: an example of ecogenetics in oncology. J. Natl. Cancer Inst. 57:3-7.
Neel, J.V. 1984. Genetics and nutrition: an evolutionary perspective. Pp. 3-16 in A. Velazquez and H. Bourges, eds. Genetics Factors in Nutrition. Academic Press, Orlando, Fla.
NIAAA (National Institute on Alcohol Abuse and Alcoholism). 1987. The Sixth Special Report to the U.S. Congress on Alcohol and Health from the Secretary of Health and Human Services. DHHS Publ. No. (ADM) 87-1519. National Institutes of Health, Public Health Service, U.S. Department of Health and Human Services. U.S. Government Printing Office, Washington, D.C. 147 pp.
Omenn, G.S. 1987. Heredity and environmental interactions. Pp. 333-341 in H.W. Goedde and D.P. Agarwal, eds. Genetics and Alcoholism. Alan R. Liss, New York.
Orkin, S.H. 1986. Reverse genetics and human disease. Cell 47:845-850.
Page, L.B. 1979. Hypertension and atherosclerosis in primative and acculturating societies. Pp. 1-12 in J.C. Hunt, ed. Hypertension Update: Mechanisms, Epidemiology, Evaluation and Management. Health Learning Systems, Bloomfield, N.J.
Poehlman, E.T., A. Tremblay, J.P. Depres, E. Fontaine, L. Perusse, G. Theriault, and C. Bouchard. 1986. Genotype-controlled changes in body composition and fat morphology following overfeeding in twins. Am. J. Clin. Nutr. 43:723-731.
Price, A., R. Cadoret, A.J. Stunkard, and B.A. Troughton. 1987. Genetic contributions to human fatness: an adoption study. Am. J. Psychiatry 144:1003-1008.
Rosenberg, L.E. 1980. Inborn errors of metabolism. Pp. 73103 in P.K. Bondy and LE. Rosenberg, eds. Metabolic Control and Disease, 8th ed. W. B. Saunders, Philadelphia.
Rosenberg, L.E. 1984. Inborn errors of nutrient metabolism: Garrod's lesson and legacies. Pp. 61-77 in A. Velazquez and H. Bourges, eds. Genetic Factors in Nutrition. Academic Press, New York.
Schaeffer, K.M., O.A. Parsons, and J.R. Yohman. 1984. Neuropsychological differences between male familial and nonfamilial alcoholics and nonalcoholics. Alcohol. Clin. Exp. Res. 8:347-351.
Sing, C.F., and J. Davignon. 1985. Role of the apolipoprotein E polymorphism in determining normal plasma lipid and lipoprotein variation. Am. J. Hum. Genet. 37:268-285.
Solomon, E., R. Voss, V. Hall, W.F. Bodmer, J.R. Jass, A.J. Jeffrey, F.C. Lucibello, I. Patel, and S.H. Rider. 1987. Chromosome 5 allele loss in human colorectal carcinomas. Nature 328:616-619.
Stanbury, J.B., J.B. Wyngaarden, D.S. Fredrickson, J.L. Goldstein, and M.S. Brown, eds. 1983. The Metabolic
Basis of Inherited Disease, 5th ed. McGraw-Hill, Trenton, N.J. 1048 pp.
Stunkard, A.J., T.I.A. Sorenson, C. Hanis, T.W. Teasdale, R. Chakraborty, W.J. Schull, and F. Schulsinger. 1986a. An adoption study of human obesity. N. Engl. J. Med. 314: 193-198.
Stunkard, A.J., T.T. Foch, and Z. Hrubec. 1986b. A twin study of human obesity. J. Am. Med. Assoc. 256:51-54.
Thurman, R.G. 1980. Ethanol elimination rate is inherited in the rat. Pp. 655-667 in R.G. Thurman, ed. Alcohol and Aldehyde Metabolizing System. IV. Advances in Experimental Medicine and Biology. Plenum Publishing Corp., New York.
Van Itallie, T.B. 1985. Health implications of overweight and obesity in the United States. Ann. Intern. Med. 103:983-988.
Velazquez, A., and H. Bourges, eds. 1984. Genetic Factors in Nutrition. Academic Press, Orlando, Fla. 434 pp.
Vogel, F., and A.G. Motulsky. 1986. Human Genetics: Problems and Approaches, 2nd ed. Springer-Verlag, Berlin. 807 pp.
Winokur, G., J. Reich, J. Rimmer, and F.N. Pitts, Jr. 1970. Alcoholism. III. Diagnosis and familial psychiatric illness in 259 alcoholic probands. Arch. Gen. Psychiatry 23:104-111.
Withers, R.F.J. 1964. Problems in the genetics of human obesity. Eugen. Rev. 56:81-90.