Abstract

Data pertinent to the issue of energy balance and body weight control obtained in a comprehensive study of diet, lifestyle and disease mortality in 65 counties (130 villages, 6500 adults) of rural China (Chen et al. 1990) were used for the analysis. After adjusting the food intake data to represent a reference male adult involved in the least physical activity and representing the same body weight, total calorie intake (40.6 kcal/kg BW) was about 30% higher in China when compared with an average adult American male (30.6% kcal/kg BW), yet the body mass index for the Chinese male was about 25% lower (20.5 vs. 25.8 kg BW/m²). Diets in rural China were low in fat (14.5% of energy), relatively low in protein (65.8 g/day), and high in fiber (33 g/day), representing a diet unusually rich in plant based foods (e.g., including about 90% of the total protein).

It is believed that the excess energy intake among the Chinese is mostly attributed to their greater physical activity, although some unknown but significant and probably difficult to measure amount could be due to increased energy expenditure associated with non-post prandial basal metabolism. This hypothesis is based, in part, on evidence from experimental animal data from this and other laboratories showing that laboratory rats fed diets comprised of substantially reduced intakes of protein consume more energy, but gain less weight and exhibit increased thermogenesis due both to enhanced metabolic body heat and to diet driven physical activity, while sharply reducing blood cholesterol concentrations and tumor development.

Introduction

Although the prevalence of obesity has been increasing at alarming rates in Western countries in recent years to levels as high as 25 to 34% of adult Americans (Kuczmarski 1992), it is now also an emerging health problem in many other countries as well (reviewed elsewhere in this conference by Popkin). Perhaps as much as $30-40 billion (Atkinson 1992) is now being expended in the United States for products and programs thought to hold promise of a possible solution. Yet, in spite of this widespread interest, control of obesity has mostly remained intractable; probably no more than 5% of those who attempt weight reduction, actually succeed (Wadden et al. 1989). It is commonly believed that those few individuals who successfully lose and maintain lower body weights are those who succeed in making substantial changes in their dietary and lifestyle practices. In contrast, gimmickry involving highly focused and quick-fix interventions are almost always unsuccessful.

Human study in rural China

Thus more thought and research are needed to better understand the chief correlates of obesity within the broader dietary and lifestyle context. One such opportunity to examine this larger context issue was our comprehensive study of diet, lifestyle, and disease mortality characteristics in rural China (Chen et al. 1990). This study was not originally designed to investigate the etiology of obesity, but instead was directed to a comprehensive investigation of the chief correlates and possible causes of chronic degenerative diseases, some of which associate with obesity.

At the start of the study, it was known that most of these diseases were much less common in China. Moreover, obesity was thought to be much less common and dietary practices were known to be very different from those of the U.S. The original motivation for this study, which included 65 counties, two villages per county, and fifty 35-64 year old adults and their families in each village, centered on findings published by the Chinese Academy of Medical Sciences (Li et al. 1981) showing that, for about a dozen different cancer sites, age-standardized mortality rates were highly localized geographically. Mortality rates ranged from as much as a few dozen to even a few hundred fold (Table 1), being far greater than the mere 1.5-2.0 fold geographic extremes observed for the US (American Cancer Society 1989).

An ideal epidemiological setting was available to investigate, with sensitivity, a multitude of diet and disease associations because 1) of the wide ranging disease mortality rates, 2) of the very stable residency patterns of the cohort subjects over their entire lifetimes (90-94% were born in the counties where they were surveyed), 3) of the consumption of locally produced food, and 4) of the remarkable consistency of dietary patterns for each survey site over time (Piazza 1986).

When the China Study was being planned, there was considerable evidence showing that the incidence of chronic degenerative diseases common in Western countries was correlated with the consumption of diets high in fat (Carroll 1975) and low in fiber (Burkitt and Trowell 1975; Watanabe et al. 1979; Committee on Diet Nutrition and Cancer 1982), that is, diets typically high in foods of animal origin. There also was a reasonable amount of experimental animal evidence on specific nutrient effects on various stages of disease development. For example, antioxidant micronutrients and various so-called 'low quality' proteins were thought and sometimes reported to retard the experimental development of cancer (Committee on Diet Nutrition and Cancer 1982). Again, this evidence favored the routine consumption of plant based diets for prevention of these chronic degenerative diseases. It was also thought by many researchers that, although individual nutrients might act under isolated experimental conditions to produce favorable effects, it was the combination of nutrients, as found in food and in dietary practices, that produced the most substantial and sustained effects. Thus, an investigation was desired which included a cohort of people experiencing a broad range of disease risks and consuming diets comprised of a range of dietary composition, especially that provided by plant derived foods. This was the opportunity available in rural China.

Across China, diets varied from those rich in plant matter on the one hand to those very rich in plant matter on the other. People tended to consume the same diets from year to year and to reside in the same region most of their lives. Diet, disease, and residency characteristics were reasonably consistent and constant, that is, over space and over time, within each survey unit. Across the whole of China, however, dietary and lifestyle characteristics (Chen et al. 1990) varied considerably, as illustrated in the widely varying social and economic conditions (Table 2).

The survey was organized and carried out under the direction of the Chinese Academy of Preventive Medicine in the Fall of 1983 and five kinds of samples and other questionnaire information were collected (Table 3). Blood, urine and food samples collected in the survey were analyzed for a variety of nutritional, viral, hormonal and toxic chemical factors between 1984-1988 while personal questionnaire information and dietary survey data were assembled and entered into computer storage. After 'cleaning' this massive amount of information during 1989-90, a total of 367 items of dietary, lifestyle and disease characteristics (130 villages, 6500 families) were judged to be reliable and were published in an 896 page monograph in 1991 (both in English and Chinese). Analysis and publication of selected portions of the data were begun in 1991 and it is mostly the findings of these published and peer reviewed papers that provides the information for this paper on possible factors concerned with energy balance and body weight sequelae. Further details on the procedure and methodology used in this survey have been published elsewhere (Chen et al. 1990).

The underlying hypothesis for the following findings states that comprehensive prevention of chronic degenerative diseases and generation of optimum health is best achieved by the consumption of a variety of good quality plant based foods. The intentional comprehensiveness of this hypothesis meant that more comprehensive methods of data analyses were required. The following therefore summarizes selected findings from this study, which mostly address associations of variables representing a variety of dietary and lifestyle exposures.

A comparison of selected dietary and other relevant characteristics for rural China and the US is shown in Table 4 (Chen et al. 1990). Average intakes of dietary fat and fiber were markedly different. The consumption of foods of animal origin, indicated by animal protein consumption, was especially low. On average, Americans consume diets containing about 10 times the concentration of animal protein than do rural Chinese. Such a difference indicates major differences in many other dietary and metabolic characteristics as well. For example, blood cholesterol and total fat intake (with mean county intakes ranging from 6% to 24% of energy) were substantially lower in rural China than in the US.

Energy (calorie) intake, per kilogram of body weight and adjusted for the physical activity equivalent of an office worker, was about 30% higher in rural China than the US, yet excess body weight appeared to be much lower, as indicated by the lower body mass index. 'Office worker' equivalent people were consuming considerably more energy but appeared to be far leaner than 'average' Americans engaged in all manner of work. Undoubtedly, much of the increased energy intake but lower body weight in rural China were attributable to their greater physical activity (e.g., most office workers ride bicycles to work) although some of this difference may also be due to the type of diet being consumed, to be discussed later.

The more comprehensive analyses and interpretation of these data were pursued in two ways. First, disease mortality rates were compared with each other to see if there were any particular tendencies for diseases with common causes to be found within similar areas of the country. Then, indices of these disease groups were examined for their associations with a large number and variety of dietary and lifestyle factors recorded in the study.

Second, we hypothesized and tested for associations among specific disease outcomes, specific diet and lifestyle causes, and specific tissue biomarkers indicative of explanatory mechanisms and biological plausibility. These associations were then assembled to determine whether they were internally consistent and supportive of this broadly based hypothesis.

The first method examines the 'outside', then peers within, while the second method examines the 'inside', then looks out. The first describes the whole 'forest' and the main 'trees' which provide the overall appearance. The second examines the trees independently and then attempts to see whether these details match the whole forest description.

In relation to the first method, two geographically localized groups of disease were found (Table 5), showing that each disease in either of these two lists is positively associated with diseases in its own list but inversely associated with diseases in the opposite list (Campbell et al. 1992). The first group (A) included diseases characteristic of developing countries while the second group (B) included diseases characteristic of the more industrialized Western countries. Such geographic aggregations of disease suggest that each disease group tends to have a common set of dietary and lifestyle causes which contribute to the etiology of each individual disease within the group. Several characteristics possibly indicative of such common causes were examined.

Characteristics which were chiefly associated with the Group B or 'Western' diseases were total blood cholesterol and plasma urea nitrogen. Plasma urea nitrogen was chiefly associated with intakes of meat (p<0.01), milk (p<0.001) and eggs (p<0.01) while blood cholesterol was directly associated with the consumption of dietary fat, meat and animal protein and inversely associated with the intakes of dietary fiber and legumes (p<0.01 for fiber, all others were p<0.05).

Total (r=0.39, p<0.01) and LDL cholesterol (r=0.40, p<0.001) were positively associated with animal protein containing diets but inversely with plant protein containing diets (Table 6). Moreover, the higher was the intake of foods of plant matter (as indicated by the fiber and legume associations), the lower was the concentration of plasma cholesterol. These findings appear to be quite remarkable because only small intakes of animal based foods were associated with significant increases in plasma cholesterol concentration and chronic degenerative diseases. The significance of this observation is drawn from the facts that 1) animal based foods comprised only 0-20% of total protein intake and 2) plasma cholesterol concentrations were already low by Western standards.

In short, disease rates were significantly associated within a range of dietary plant food composition that suggested an absence of a disease prevention threshold. That is, the closer a diet is to an all-plant foods diet, the greater will be the reduction in the rates of these diseases.

The social and economic conditions associated with the development of these two groups of disease are reasonably well-known. Degenerative diseases of Western countries, which are commonly known as 'diseases of affluence' but which are, in our view, more appropriately characterized as 'diseases of nutritional extravagance', tend to occur in areas where urbanization, industrialization, and wealth converge. As a society begins to acquire additional capital and resources, people appear to quickly begin to consume nutritionally richer diets, particularly diets richer in added fat and sugar and foods of animal origin. This phenomenon has characterized the industrialization of countries around the world throughout history. Wealth seems to catalyze the emergence of Western type diseases perhaps because instant gratification with dietary and lifestyle experiences become affordable and are desired.

The second method of data interpretation provided further insight into these more generalized diet and disease associations by investigating the more specific associations, then to be followed by an aggregation of these findings in reference to the general hypothesis under investigation (i.e., describing the forest by investigating individual trees). A broad variety of associations have thus far been investigated, as shown in the following selection of findings.

Breast cancer was directly associated with increasing intakes of dietary fat and higher concentrations of blood cholesterol (Key et al. 1990; Marshall et al. 1991; Wang et al. 1991) and was best predicted by plasma testosterone concentrations associated, in turn, with diets higher in fat and foods of animal origin. Plasma estradiol concentrations, although too variable to be associated with breast cancer in this study, nonetheless were 50-60% lower than those of comparable British women, an observation which is consistent with the lower levels of breast cancer in China. Probably the most interesting observation was the substantial difference in age at menarche, averaging 17 years in rural China and 12 years in the US (mean age at menarche for Chinese counties ranged as high as 19 years). It is possible that nutritionally rich diets, which are well known to increase childhood growth rates, may cause sexual maturation (menarche) to occur earlier (van Wieringen 1986; Tretli and Gaard 1996). Many studies also have shown that the earlier is the age of menarche, the greater is the risk for breast cancer later in life, as reviewed by Kelsey (Kelsey et al. 1993). In this context, the fastest rate of childhood growth may not be the healthiest.

Body height attained during adulthood was positively associated with increasing intakes of plant protein (Chen et al. 1990), suggesting that the genetic potential for body height can be accomplished simply by consuming diets with adequate intakes of plant-based foods. That is, it is not necessary to consume animal-based protein containing foods to reach our genetically determined body height, as most professional and lay people around the world have assumed. Consuming adequate quantities of plant-based foods to reach our ultimate body size also has the added advantage of minimizing risk for the degenerative diseases. This finding is consistent with the observations on childhood growth rates of vegetarian children regularly consuming a varied diet of good quality food (Sabate et al. 1990; Sabate et al. 1992). Although growth rates may be slightly but insignificantly lower for some vegan and vegetarian children, this probably represents a considerable health advantage, certainly insofar as breast cancer risk later in life is concerned.

The higher were the plasma levels of vitamin C and beta-carotene, the lower were the rates of several cancers (Gao 1990; Chen et al. 1992). These antioxidant vitamins are provided almost entirely by plant-based foods.

Dietary fiber is infinitely complex and its intake can be assessed in many different ways. Associations of multiple fiber constituents with cancers of the large bowel (Campbell et al. 1990; Chen et al. 1990) consistently showed lower cancer rates with increased intakes of these fibrous foods, although these associations were only marginally significant, perhaps because these fiber intakes were unusually high when compared with Western practices.

Although certain cancers commonly found in the developing countries, such as with liver and stomach cancers, may be crudely associated with the consumption of plant based foods, this does not infer causality. One likely reason that these cancers are more common in these relatively poor countries is because of increased exposure to non-nutrient factors often associated with impoverishment but generally required for disease onset. Thus, there are much higher sub-sets of the population who are susceptible to these diseases. With liver cancer, it is chronic infection to hepatitis B (Campbell et al. 1990), and C (Okuda 1991) viruses, and with stomach cancer, it is the extensive use of non- refrigerated, highly salted and fermented foods, as reviewed by a recent report of diet and cancer worldwide (Expert Panel 1997). Such foods enhance the development of chronic stomach 'ulcers' associated with a bacterial organism, Helicobacter pylori (Forman et al. 1990). Among people who are predisposed, evidence from this study shows that increasing intakes of foods of plant origin actually reduces the likelihood of disease, both for liver cancer (Campbell et al. 1990) and for stomach cancer (Kneller et al. 1992).

Dietary associations with various kinds of cardiovascular disease have also been examined from multiple perspectives (Campbell et al. 1998). Apolipoprotein B, an index of 'bad' cholesterol, was increased with increasing rates of disease. In turn, this cholesterol level was associated with increasing intakes of meat and animal protein but was inversely associated with increasing intakes of legumes, 'light' vegetables, cellulose and plant protein. Again, such findings emphasize the health value of plant-based foods.

The dietary and nutritional inferences of these many associations, both general and specific, are many. A wide variety of specific diet-disease relationships in this study indicate considerable health advantage provided by plant-based diets. The likelihood of there being so many associations within this particular range of diet composition pointing toward a plant based diet also appears to be highly statistically significant; that is, virtually little or no associations have thus far been found to indicate otherwise. Moreover, the fact that there are a plethora of explanatory mechanisms and a virtually unlimited number of possible causes makes this suggestion even more plausible.

Consequences of shift in energy balance

The principal data from this study related to energy balance and its sequelae relate to the higher energy intake and the lower body mass, even after adjustment of the Chinese intakes to represent individuals having the least physical activity. The inability of the excess energy consumed by these individuals to create excess body weight, as expected, is likely due to their greater physical activity. However, it is also possible that some of this effect could result from increased energy expenditure as body heat (i.e., thermogenesis) attributed to the consumption of this low protein, low fat diet. We suggest, for example, that a small but relatively unmeasurable increase in energy expenditure can spare an otherwise significant increase in energy retention associated with a relatively large gain in body weight.

Elsewhere in this conference, Pi-Sunyer offered the idea that a small but unmeasurable increase in energy intake, say, of only 2% of a 2000 kcal/day diet (i.e., 40 kcal/day) could readily give rise to a body weight gain of 5 pounds per year (this assumes that 1 g of body weight gain would require 6.4 kcal of unmeasurable excess energy intake). Many years ago, Hegsted (Hegsted 1976) also made essentially the same point, when he concluded that "An excess of intake over expenditure of 50 kcal per day had the potential for producing 6 to 7 g of adipose tissue per day or 2 kg per year. Yet in an individual consuming 2500 kcals per day, 50 kcals represent only 2 percent of the total intake." These estimates of Pi Sunyer and Hegsted are similar. Pi-Sunyer assumes that 6.4 kcal is required for 1 g of accumulation of whole tissue, while Hegsted assumes that 9 kcal is required for accumulation of adipose tissue only. The point made by both investigators is essentially the same, namely that the amount of excess energy intake required for a significant accumulation of body weight is too small to be reliably measured.

It is also possible that, instead of biologically meaningful energy intakes being measured with difficulty, as suggested by Pi-Sunyer, "there are equally severe problems in estimating...caloric expenditure", as suggested by Hegsted (Hegsted 1976). This is related to the well established idea (Miller and Payne 1962; Samonds and Fleagle 1973) that excess body weight is also related to the relative proportions of energy intake 'retained' as body weight or, conversely, expended as body heat, either as increased physical activity and/or increased resting metabolic rate (i.e., thermogenesis^a). A relatively small but virtually unmeasurable shift of only 40-50 kcal/day of metabolic energy away from body heat expenditure to the formation of body weight would produce the same body weight gain as that suggested by Pi-Sunyer.

Here, we turn our attention to the experimental animal literature. There is a well established literature on animal production efficiency showing that feeding diets sufficiently low in protein will lead to less efficient energy utilization and slower body weight gain (Miller and Payne 1962). Although a diversion of energy away from weight gain to body heat may be attributed to multiple biochemical mechanisms (Freake and Oppenheimer 1995), that due to brown adipose tissue metabolism may by especially significant, at least in experimental animals (Miller and Payne 1962; Rothwell and Stock 1981). Although it is uncertain how significant is brown adipose tissue metabolism for humans, it has not, in our opinion, been ruled out, especially that very small amount which would otherwise be required for producing biologically meaningful but difficult to measure gain in body weight.

We make this point in part because of findings from our own experimental animal studies. We have for many years fed rats diets containing varying amounts of protein (mostly as casein) for the purpose of exploring its effect on the development of chemically induced tumors (Appleton and Campbell 1983; Appleton and Campbell 1983; Dunaif and Campbell 1987; Youngman 1990), that is, as aflatoxin-induced hepatocellular carcinoma. Routinely, diets containing 5% casein prevent virtually all tumor development and substantially reduce circulating cholesterol levels when compared with diets containing 15-20% casein. In addition, the 5% casein fed animals, while surviving longer than all of the 20% casein fed animals, also routinely consume slightly but significantly more energy yet gain body weight more slowly, thus reflecting lower efficiency of energy utilization but higher expenditure of energy as body heat. Increased consumption of oxygen and increased quantities of brown adipose tissue have also been observed in these low protein fed animals (Horio et al. 1991), thus being consistent with other studies (Rothwell et al. 1983). Most remarkably, the low protein fed rats increase rather substantially their voluntary physical activity, as measured in exercise wheels (Krieger 1988). Thus, thermogenesis is increased in the low protein fed animals both by greater brown adipose tissue metabolism and by greater voluntary physical activity.

Hawrylewicz and co-workers (Huang et al. 1982; Hawrylewicz 1986; Hawrylewicz et al. 1986) have reported similar observations for rats fed modestly low protein diets and induced with a mammary carcinogen. Increased energy consumption, decreased body weight gain, inhibition of tumor development and decreased circulating estrogen levels each have been reported.

In short, decreased intake of dietary casein in both of these well studied experimental animal models is analogous to decreased intake of animal based foods by rural Chinese. Moreover, the interpretation of these findings is essentially the same for both species. A low protein diet, especially one low in the more efficiently utilized animal based protein, is associated with lower efficiency of energy retention to form the excess body weight, that is, more energy is dispensed either through an increase in thermogenesis or through enhanced physical activity. And most importantly, the shift of energy expenditure away from deposition as body fat toward body heat is much too small to be reliably measured, yet this amount is more than enough to spare significant gains in body weight. Moreover, the possibility that adaptation to new steady states of thermogenesis (Waterlow 1986) can be prolonged makes even more difficult its detection.

It should also be noted that there is yet another diet induced shift in energy balance that could readily account for the less efficient retention of energy by the rural Chinese. This is that which is associated with the low fat intakes, a situation somewhat equivalent to the so-called very low calorie diets studied in the Western literature. Forbes and co-workers long ago (Forbes et al. 1946; Forbes et al. 1946) showed with young growing rats that low fat diets produced lower energy efficiency than did high fat diets. Indeed, they concluded that the caloric value of fat, upon metabolism, was likely closer to 10 kcal/g than the Atwater figures of 9 kcal/g, an observation that was similarly advocated by Donato and Hegsted (Donato andHegsted 1985) who suggested a value of 11 kcal/g.

In conclusion, in rural China where the prevalence of obesity is uncommon, diets low in fat and total protein and high in fiber were found to be associated with greater energy intakes but lower body weights. Using experimental rat data from this and other laboratories for developing biologically plausible hypotheses, we suggest that healthful, non-obese body weights can be controlled through the regular consumption of low fat, low protein diets which shunt a relatively small amount of energy expenditure away from body fat deposition to body heat. The diet induced shift in energy disposition, attributed to relatively small and difficult to detect increases in basal metabolism and physical activity, would still be sufficient to account for otherwise substantial increases in body weight.

References

American Cancer Society (1989). Cancer facts & figures-1989. Atlanta, GA, American Cancer Society, Inc.

Appleton, B. S. and T. C. Campbell (1983). Dietary protein intervention during the post-dosing phase of aflatoxin B1-induced hepatic preneoplastic lesion development. J. Natl. Cancer Inst. 70, 547-549.

Appleton, B. S. and T. C. Campbell (1983). Effect of high and low dietary protein on the dosing and postdosing periods of aflatoxin B1-induced hepatic preneoplastic lesion development in the rat. Cancer Res. 43, 2150-2154.

Atkinson, R. L. (1992). Editorial: Treatment of obesity. Nutr. Revs. 50, 338-345.

Burkitt, D. P. and H. C. Trowell (1975). Refined carbohydrate foods and disease: some implications of dietary fibre. London, Academic Press.

Campbell, T., B. Parpia, et al. (1998). The contextual framework surrounding the causes of coronary heart disease. Am. J. Cardiology 82, in press.

Campbell, T. C., J. Chen, et al. (1992). China: from diseases of poverty to diseases of affluence. Policy implications of the epidemiological transition. Ecol. Food Nutr. 27, 133-144.

Campbell, T. C., J. Chen, et al. (1990). Non-association of aflatoxin with primary liver cancer in a cross-sectional ecologic survey in the People's Republic of China. Cancer Res. 50, 6882-6893.

Campbell, T. C., G. Wang, et al. (1990). Dietary fiber intake and colon cancer mortality in The People's Republic of China. In Dietary Fiber. D. Kritchevsky, C. Bonfield and J. W. Anderson pp. 473-480.Plenum Publishing Corporation New York, NY.

Carroll, K. K. (1975). Experimental evidence of dietary factors and hormone-dependent cancers. Cancer Res. 35, 3374-3383.

Chen, J., T. C. Campbell, et al. (1990). Diet, life-style and mortality in China. A study of the characteristics of 65 Chinese counties. Oxford, UK; Ithaca, NY; Beijing, PRC, Oxford University Press; Cornell University Press; People's Medical Publishing House.

Chen, J., C. Geissler, et al. (1992). Antioxidant status and cancer mortality in China. Int. J. Epidemiol. 21, 625-635.

Committee on Diet Nutrition and Cancer (1982). Diet, nutrition and cancer. Washington, D.C., National Academy Press.

Donato, K. and D. M. Hegsted (1985). Proc. Natl. Acad. Sci, USA 82, 4866-4870.

Dunaif, G. E. and T. C. Campbell (1987). Relative contribution of dietary protein level and Aflatoxin B1 dose in generation of presumptive preneoplastic foci in rat liver. J. Natl. Cancer Inst. 78, 365-369.

Expert Panel (1997). Food, nutrition and the prevention of cancer:a global perspective. Washington, DC, World Cancer Research Fund and The American Institute for Cancer Research.

Forbes, E. B., R. W. Swift, et al. (1946). J. Nutr. 32, 387-396.

Forbes, E. B., R. W. Swift, et al. (1946). J. Nutr. 32, 397-403.

Forman, D., F. Sitas, et al. (1990). Geographic association of Helicobacter pylori antibody prevalence and gastric cancer mortality in rural China. Int. J. Cancer 46, 608-611.

Freake, H. C. and J. H. Oppenheimer (1995). Thermogenesis and thyroid function. In Ann. Rev. Nutr. D. B. McCormick, D. M. Bier and A. G. Goodridge. 15 pp. 263-291.Annual Reviews, Inc. Palo Alto, CA.

Gao (1990). Nutr. Cancer 13, 121.

Hawrylewicz, E. J. (1986). Fat-protein interaction, defined 2-generation studies. In Dietary fat and cancer. C. Ip, D. F. Birt, A. E. Rogers and C. Mettlin pp. 403-434.Alan R. Liss, Inc. New York.

Hawrylewicz, E. J., H. H. Huang, et al. (1986). Dietary protein enhancement of N-nitroso-methylurea-induced mammary carcinogenesis, and their effect on hormone regulation in rats. Cancer Res. 46, 4395-4399.

Hegsted, D. M. (1976). Energy needs and energy utilization. In Present Knowledge in Nutrition, Fourth Edition. D. M. Hegsted, C. O. Chichester, W. J. Darbyet al pp. 1-9.The Nutrition Foundation, Inc. New York.

Horio, F., L. D. Youngman, et al. (1991). Thermogenesis, low-protein diets, and decreased development of AFB1-induced preneoplastic foci in rat liver. Nutr. Cancer 16, 31-41.

Huang, H. H., E. J. Hawrylewicz, et al. (1982). Effect of protein diet on release of prolactin and ovarian steroids in female rats. Nutr. Rpts. Int. 26, 807-820.

Kelsey, J. L., M. D. Gammon, et al. (1993). Reproductive factors and breast cancer. Epidemiol. Revs. 15, 36-47.

Key, T. J. A., J. Chen, et al. (1990). Sex hormones in women in rural China and in Britain. Brit. J. Cancer 62, 631-636.

Kneller, R. W., W.-D. Guo, et al. (1992). Risk factors for stomach cancer in sixty-five Chinese counties. Cancer Epi. Biomarkers Prev. 1, 113-118.

Krieger, E. (1988). Increased voluntary exercise by Fisher 344 rats fed low protein diets. Ithaca, NY, Cornell University.

Kuczmarski, R. J. (1992). Prevalence of overweight and weight gain in the United States. Am. J. Clin. Nutr. 55, 495S-502S.

Li, J.-Y., B.-Q. Liu, et al. (1981). Atlas of cancer mortality in the People's Republic of China. An aid for cancer control and research. Int. J. Epid. 10, 127-133.

Marshall, J. R., Y. Qu, et al. (1991). Additional ecologic evidence: lipids and breast cancer mortality among women age 55 and over in China. Europ. J. Cancer 28A, 1720-1727.

Miller, D. S. and P. R. Payne (1962). Weight maintenance and food intake. J. Nutr. 78, 255-262.

Okuda, K. (1991). Hepatitis C virus and hepatocellular carcinoma. In Etiology, pathology, and treatment of hepatocellular carcinoma in North America. E. Tabor, A. M. Di Bisceglie and R. H. Purcell pp. 119-126.Gulf Publishing Company Houston, TX.

Piazza, A. (1986). Food consumption and nutritional status in the People's Republic of China. London, Westview Press.

Rothwell, N. J. and M. J. Stock (1981). Regulation of energy balance. Ann. Rev. Nutr. 1, 235-256.

Rothwell, N. J., M. J. Stock, et al. (1983). Mechanisms of thermogenesis induced by low protein diets. Metabolism 32, 257-261.

Sabate, J., K. D. Lindsted, et al. (1990). Attained height of lacto-ovo vegetarian children and adolescents. Europ. J. Clin. Nutr. 45, 51-58.

Sabate, J., M. C. Llorca, et al. (1992). Lower height of lacto-ovovegetarian girls at preadolescence: an indicator of physical maturation delay? J. Am. Diet. Assoc. 92, 1263-1264.

Samonds, K. W. and J. Fleagle (1973). The onset of protein and calorie deficiency in the young Cebus monkey. Fed. Proc. 32, 901 only.

Tretli, S. and M. Gaard (1996). Lifestyle changes during adolescence and risk of breast cancer: an ecologic study of the effect of World War II in Norway. Cancer Causes Control 7, 507-512.

van Wieringen, J. (1986). Secular growth changes. In Human growth: a comprehensive treatise. F. Faulkner and J. Tanner pp. 307-331.Plenum Publishing Corp. New York.

Wadden, T. A., J. A. Sternberg, et al. (1989). Treatment of obesity by very low calorie diet, behavior therapy, and their combination: a five year prospective. Int. J. Obesity 13(Suppl.2), 39-46.

Wang, D. Y., T. J. A. Key, et al. (1991). Serum hormone levels in British and rural Chinese females. Brit. Cancer Res. Treatment 18, S41-S45.

Watanabe, R. B., B. S. Reddy, et al. (1979). Effect of dietary alfalfa, pectin, and wheat bran on azoxymethane- or methylnitrosourea-induced colon carcinogenesis in F344 rats. J. Natl. Cancer Inst. 63, 141-145.

Waterlow, J. C. (1986). Metabolic adaptation to low intakes of energy and protein. Ann. Rev. Nutr. 6, 495-526.

Youngman, L. D. (1990). The growth and development of aflatoxin B1-induced preneoplastic lesions, tumors, metastasis, and spontaneous tumors as they are influenced by dietary protein level, type, and intervention. Ithaca, NY, Cornell University, Ph.D. Thesis.