Chapter 8 : Health, obesity and energy values of dietary fat

Contents - Previous - Next

Health risks associated with obesity
Causes of obesity
Energy values of fat
Fat substitutes

Overweight is defined as an increase above a standard in body weight related to height. It is often measured by the Quetelet or Body Mass Index (BMI) which is weight in kilograms divided by the square of stature in metres (BMI = weight/height2). Obesity is also defined as an abnormally high percentage of body fat. In males, normal body fat represents 12-20 percent of body weight. Body fat represents 20-30 percent of body weight in normal females.

Weight standards for a population can be established by dividing a large sample of the population according to the normal distribution of their body weights in relation to height. The inherent difficulty with this approach is that there is an underlying assumption that average weights are healthy weights. An alternative approach to determining standards for healthy weights is to use body weights which are associated with the lowest overall risk to health. In several prospective studies, the lowest death rate was associated with a BMI close to 22 (Manson et al., 1987; Garrison and Kannel, 1993).

Longitudinal tracking of weight. Several epidemiological studies have examined the weight status of various populations over time (Noppa and Hallstrom, 1981; Borkan et al., 1986; Williamson et al., 1990) and they show a progressive rise in BMI in most populations. However, the weight status of an individual from infancy and childhood through adolescence and into adult life often follows an uneven track. (Bradden et al., 1986; Charney et al., 1976; Khouryet all, 1983; Mossberg, 1989; Zacketal., 1979).

A few studies (Garn et al., 1986; Johnston and Mack, 1978; Melbin and Vuille, 1976) have calculated the relative risk of being in the top weight category as an adult based on weight status in childhood or adolescence. When the lightest youngsters were compared with the heaviest ones, the heaviest youngsters were 1.6 to 2.5 times more likely to be overweight as adults. However, the Harvard Longitudinal Studies of Child Health and Development 50 year investigation found that the BMI of females during childhood had essentially no correlation with their BMIs when they reached middle-age (Casey et al., 1992). There was a better correlation between the BMIs of adolescent females and those of women at age 50 years, although this was still low. In females, the failure to show an effect of overweight in adolescence on mortality in adults may be due to this low correlation. In men, on the other hand, the correlation of BMI in childhood or adolescence with BMI at age 50 years was better, but the tracking of childhood and adolescent weight into adult life is still of a low order (Borkan et al., 1986).

Children from families where one or both parents are overweight are at higher risk of adult obesity. If these children are notably obese during their school years, a preventive programme may be advisable. Thus, ongoing surveillance of children and adolescents will help identify those who are at risk and may be helped by preventive service.

Health risks associated with obesity

Overweight and fat distribution are useful predictors of premature mortality and the risks of heart disease, hypertension, non-insulin-dependent diabetes mellitus, gallbladder disease and some cancers. However, if body fat per se were the major risk factor associated with premature mortality, obese women might be expected to have shorter life expectancies than obese men. This is generally not the case, and it is now recognized that it is fat distribution, especially an increase in abdominal and visceral fat, that is a predictor of the health risk related to obesity. For example, weight gain of less than 5 kg in women during adult life may carry little extra risk, particularly if the added weight is located in the femoral region. For most men, any weight gain after the age of 20 years increases risk because this fat is usually deposited as abdominal and visceral fat.

In epidemiology, the relationship between data on BMI and a given disease risk is shown to be curvilinear and this curve is often described as J- or U- shaped. This means that mortality and morbidity tend to increase as the BMI increases above 25 or drops below 18.5.

Life insurance statistics from the United States show that excess weight is associated with higher mortality rates. Based on 1979 data (Society of Actuaries and Association of Life Insurance and Medical Directors of America, 1980), a body weight which is 10 percent above average weight is accompanied by an 11 percent increase in excess mortality for men and a 7 percent increase for women. If body weight is 20 percent above average, the excess mortality rises to 20 percent for men and 10 percent for women. Conversely, the minimum mortality rates for both men and women occur among individuals who are approximately 10 percent below average weight.

More than six studies have been published showing that central adiposity is positively correlated with increased mortality as well as the risk for developing cardiovascular disease, diabetes mellitus and stroke (Lapidus et al., 1984; Larsson et al., 1984; Donahue et al., 1987; Ducimetiere, Richard and Cambien, 1986; Stokes, Garrison and Kannel, 1985). Among the highest quintile for central body fat, the relative risk of myocardial infarction was 8.2 times higher than that for the lowest quintile. For stroke and mortality in general, the relative risk was 3.8 and 2.8 times higher for individuals in the highest quintile compared to those in the lowest quintile.

Losing and regaining weight, so-called weight cycling, may also be hazardous. Data from the Chicago Gas and Electric Company Study (Hamm, Shekelle and Stamler, 1989) showed that those people who gained and lost weight had a significantly higher risk of death from cardiovascular disease than the group of individuals with no change in weight. More recently, data from the Framingham study were used to show that significant changes in weight were associated with a higher likelihood of mortality (Lissner et al., 1991). Caution . in interpreting data on weight cycling as a cause of detrimental changes is urged.

Obesity must modify some intermediate mechanism such as cardiac function, the metabolism of lipids or glucose to produce death or disease. However, even though severe overweight generally increases the risk of death, especially sudden death, in many studies it could not be seen as an independent variable. Two major problems plague the interpretation of studies in terms of obesity as an independent factor (Manson et al., 1987). First, many studies fail to separate smokers from non-smokers. Since smokers as a group tend to have lower body weights and higher mortality rates, including them in a study population influences the death rates and confounds the assignment of effects to body weight per se. Second, early mortality may bias the interpretation of weight status on life expectancy. For example, individuals who have lost or are losing weight at the time of an initial survey may die and thus overemphasise the effect of low body weights as a cause of higher mortality. The failure to identify obesity as an independent risk factor has led many to suggest that it is unimportant.

Diabetes mellitus. More than 90 percent of all diabetics have non-insulin dependent diabetes mellitus (NIDDM), and excessive weight gain and overweight are the major nutritional factors which increase the risk of developing NIDDM. The importance of body weight is demonstrated by the low level of NIDDM during World War II and during periods of famine. NIDDM is almost non-existent in individuals with a BMI of 20 or less. A BMI of 35 has an 8-fold increase in risk of NIDDM when compared with a BMI of 25.

Other types of diabetes mellitus are insulin-dependent (IDDM), gestational diabetes and a very rare form, malnutrition-related diabetes mellitus. IDDM is an immunologic disease which destroys the insulin-producing cells. Patients with this disease require insulin for survival. The nature of the autoimmune stimulus to pancreatic destruction is unknown. The role of diet in the development of this disease is not established.

The San Luis Valley study in the USA provided some evidence to support the hypothesis that a high fat, low-carbohydrate diet increases the risk for onset of non-insulin-dependent diabetes (Marshall, Hamman and Baxter, 1991). This involved cross-sectional analyses of individuals without a prior history of diabetes. A 24-hour diet recall preceded a glucose tolerance test. A high intake of fat appeared to be associated with impaired glucose tolerance, however, this may be the result of weight gain.

Other evidence comes from groups who changed their dietary patterns. These studies suggest that higher intakes of dietary fat and/or the weight gain which is associated with such increases, may be a factor in the onset of non-insulin-dependent diabetes. The Australian Aborigines were prone to the disease when they made a transitio,n from their traditional manner of living to an urban life-style (O'Dea, White and Sinclair, 1988). The mild elevation in plasma triglycerides and fasting insulin levels were consistent with insulin resistance. A group of Bangladeshis who had emigrated to the UK were found to have low plasma cholesterol concentrations, enhanced insulin resistance and three times the usual prevalence of diabetes, as well as a high rate of morbidity and mortality from coronary heart disease (McKeigue et al., 1988). A deterioration in carbohydrate tolerance and changes in lipoprotein patterns of the Pima Indians and Caucasians also point toward problems with the modern, high fat diet and obesity (Swinburn et al., 1991).

Ethnic differences in the C-peptide response relative to the insulin response in testing for glucose tolerance led to the tentative conclusion that dietary fat may be involved in determining insulin secretion and its clearance in the liver (Cruickshank et al., 1991). There is still more speculation than evidence to link dietary fat with non-insulin-dependent diabetes. At the present time, the general rationale for modifying dietary fat intakes has been to reduce the risk of coronary heart disease in diabetics.

Causes of obesity

Nutrient balance of fat storage. Obesity is always a chronic failure to balance nutrient intakes with expenditure (oxidation) (Bray, York and Fisler, 1989). There are various causes of obesity. At one extreme, obesity may be due simply to excess food (energy) intake in relation to energy requirements. In these cases, factors of heredity play an important role in the genesis of obesity which can develop even when the diet is composed primarily of carbohydrates. At the other extreme, there are those types of obesity where dietary composition, particularly a high fat intake, is central to its development. Any of these types of dietary obesity can be controlled by changing the composition of diets, by restraining the intake of food, or by increasing nutrient oxidation.

While overall energy balance is central to obesity, the concept of macronutrient balance is also useful for understanding the factors affecting excess weight gain and loss. In a normal adult, the daily energy intake from carbohydrates would be between 50 and 100 percent of the total carbohydrate stores, in the body. In contrast, the protein intake would be a little over 1 percent of total stores while the fat intake would be considerably less than 1 percent of that stored in the body. Glycogen (the storage form of carbohydrate) metabolism is closely regulated and, depending on the balance between carbohydrate intake and oxidation, there can be wide fluctuations in the carbohydrate stores over even short periods of time. This is not the case with protein and fat stores for which it takes considerably longer to affect appreciable change.

The process of regulating nutrient balance is complex. The feedback model for nutrient balance is described as having four components (Bray, 1987). The first is the "controlled system" which consists of the intake, digestion, absorption, storage and metabolism of the nutrients in food. The second is the "controller" located in the brain, and the third is comprised of the feedback signals which tell the controller about the state of the controlled system. Pinally, there are the mechanisms that modulate nutrient intake and energy expenditure.

In general, there are effective feedback and control mechanisms for regulating and balancing carbohydrate intake and oxidation (Flats, 1988). It is more difficult to become obese by eating a very high carbohydrate diet than by eating a high fat diet for a number of reasons. First, the quantity (volume) of high carbohydrate/high fibre foods which is required is much larger than that needed with a high fat diet. Second, the storage capacity for carbohydrates is limited. Third, the biochemical pathways for conversion of carbohydrate to fat are limiting and energetically expensive and are virtually negligible under normal dietary conditions in humans. Finally, the intake of carbohydrate stimulates carbohydrate oxidation so that carbohydrate balance is maintained once glycogen stores are filled, a process also stimulated by carbohydrate consumption.

Protein balance is also well-controlled. Protein stores rise gradually and only in response to stimuli other than increased protein intake. Protein which is consumed in excess of what is needed for tissue building and repair as well as enzyme formation, is converted to carbohydrate. Positive protein balance can contribute to overall energy balance in the same way that positive carbohydrate balance does.

Chronic imbalance between fat intake and fat oxidation can lead to changes in fat stores of adipose tissues. Avoiding storage of the fat consumed in a high fat diet requires that the dietary fats be oxidized. In clinical studies low rates of fat oxidation in the basal state predict an increase in body weight (Zurlo et al., 1990). It is only when the oxidation of fat equals the intake of fat that a stable body weight can be achieved.

While carbohydrate and protein oxidation are affected by carbohydrate and protein consumption, fat oxidation is not affected by fat intake and the day-to-day relationship between fat balance and fat intake is weak (Flats, 1988). Fat oxidation is related most strongly to energy balance (with a negative energy balance promoting fat oxidation) and also to the degree of body fat (Schulz et al., 1992; Zurlo et al., 1990). Creating a negative energy balance through exercise or dietary restriction can effectively increase fat oxidation, as can reducing the fat content of the diet. However, as weight is lost, fat oxidation also tends to decline. To prevent an individual from regaining weight which has been lost, the intake of fat must be reduced by approximately 20 g/day for each 10 kg of fat lost (Schulz et al., 1992).

There are major differences between individuals in their capacity to increase the oxidation of fat after beginning a high fat diet (Zurlo et al., 1990). While much of this difference is genetic, physical training can increase oxidation of fatty acids by muscle and reduce the tendency to gain weight. For this reason, adequate physical activity should be part of any weight control programme.

Some papers have shown a weak positive correlation between fat intake and body weight (WHO, 1990; Romieu et al., 1988, Miller et al., 1990) but others have not. Access to a diet with more than 30 percent energy from fat produces obesity in many animals. It was found that women who have lowered fat intakes have reduced body weight providing further evidence of a positive association between fat intake and body weight. However, in longer trials the effect has been small, probably because compliance during long studies fails (Sheppard, Kristal and Kushi, 1991; Lissner et al., 1991, Lee-Han et al., 1988).

Diet Caloric Density. It is well-recognized that energy per 100 grams of diet, referred to as caloric density, increases as fat content rises. Low-fat diets have resulted in weight loss in short-term studies, however, in long-term trials, low-fat diets in premenopausal women resulted in the consumption of an additional 19 percent of energy to maintain weight (Prewitt, 1991). The real long-term change is unknown because food was supplied to the subjects in these studies. In free-living persons, reinforcement and motivation are needed for adherence to a low-fat diet.

From the data available, it is clear that obesity is a chronic problem for which the cure is very elusive. Weight loss usually occurs during effective treatment; however, when treatment is terminated weight regain is common. This is consistent with the course of most chronic problems, and clearly indicates that prevention is the preferred method of controlling obesity.

Energy values of fat

Fatty acids are the dietary components that account for the highest amount of energy liberated during oxidation of lipids. Glycerol, to which most fatty acids are esterified, is a 3-carbon molecule with 3 oxygen atoms. Although glycerol is 10 percent by weight of triacylglycerol, it accounts for only 5 percent of the energy. The fatty acids of food impart the high energy value that is of nutritional interest.

The traditional factor for calculating the amount of dietary fat is 9 kcal (37.7 kj)/g compared to 4 kcal (16.7 kj)/g for carbohydrate and protein (FAO/WHO, 1978). Onginally these values were proposed by Atwater, and are based on the amounts of energy liberated when these macronutrients are metabolically oxidized, allowing for incomplete intestinal absorption.

The factor of 8.37 has been used for many decades to convert grams of fat in cereals, fruits and vegetables to calories (Merrill and Watt, l9SS). The precision in this 3-digit number is unwarranted, particularly upon examination of its derivation.

At the turn of the century, Atwater proposed factors for estimating the energy value of nutrients. The heat of combustion for triglycerides was applied to extracts of various foods, but for cereals, fruits and vegetables the "assumed or calculated" value was given as 9.30 calories per gram. Merrill and Watt (1955) copied this figure from a table created in 1900 by Atwater and Bryant. Merrill and Watt wrote: "For fat as it occurs in cereals and other plant sources Atwater assumed the apparent digestibility to be 90 percent and we have continued this practice. The energy factor for fat in plant foods is therefore, assumed to be 8.37 calories per gram."

Since Atwater had used a coefficient of digestibility of 95 percent for butter, it was assigned also to the separated fats of plant origin. The calculation for fats and oils as such became 95 percent of 9.30 or 8.84 calories per gram. These factors have a long history of many assumptions. The corrections for digestibility are inappropriate for the dietary oils and fats commonly consumed, and the factor of 9 for converting grams of all dietary fat to calories is more suitable and offers consistency.

Fat substitutes

Advice to reduce fat and energy consumption has led to the production of foods with a lower content of fat and to the development of fat substitutes. There are two principal approaches to the replacement of dietary fat. The first involves hydratable carbohydrates and proteins with the mouthfeel of fats. The second includes non-absorbed synthetic substances with the physical properties and technical function of fat within foods.

Protein and carbohydrate materials can be produced in forms that binds up to three times their weight in water and that create particles that imitate the mouthfeel of fat. When the hydrated material replaces fat in food, the caloric reduction is from 9 kcal/g to 1 to 2 kcal/g. The carbohydrates used are the low molecular weight starches, dextrins, maltodextrins and gums (xanthan for example), the proteins for fat replacers come primarily from milk and egg white. Since these substances have not been altered chemically, they are digested, absorbed and metabolized like ordinary nutrients (Vanderveen and Glinsmann, 1992).

The other group of fat substitutes, which has not yet reached the market, is produced by chemical synthesis from a wide range of possible structures having different digestibilities and effects on gastrointestinal function. Such material replaces fat on a gram to gram basis. The concern about some non-absorbable lipophilic substances such as fatty acid esters of sugars revolves around impaired absorption of fat-soluble compounds, both vitamins and drugs. For example, the prolonged use of sucrose polyester had a negative effect on the body stores of d-tocopherol (Jandacek, 1991). Individual assessments for safety are required for each type of synthesized fat substitute (Borzelleca, 1992; Vanderveen and Glinsmann, 1992).

The value of these kinds of materials for reducing either the total energy consumption, as an aid in the treatment or prevention of obesity, or as modifiers of a kind of dietary fat remains to be investigated. The assessment of fat substitutes must include the effect on the total diet and the redistribution of macro- and micronutrients (Rolls and Shide, 1992).

Extrapolation from animal studies may be inappropriate since human beings can choose replacements from a wide variety of foods.

Special attention must be given to the effect of fat substitutes on the consumption of micronutrients, such as vitamins A, D and E, since certain dietary fats are often important sources of these nutrients and the amount of dietary fat may be important in the utilization of such materials.


Obesity and excess visceral fat are conditions that predispose individuals to chronic diseases. Excess body fat is a major risk-factor of morbidity and mortality in most affluent countries and, increasingly, in many developing countries.

High intakes of dietary fat are associated with increasing levels of obesity in animals. For humans, the relationship between the rates of fat intake and oxidation is clear, however, the association between fat intake as a proportion of energy and obesity is weak.

Prevention of obesity involves early recognition and surveillance, regular physical activity and, for adults, a restrained pattern of eating.

Contents - Previous - Next