Risk of CHD

Low fat, high carbohydrate diets, compared to higher fat intakes, can induce a lipoprotein pattern called the atherogenic lipoprotein pheno-type (Krauss, 2001) or atherogenic dyslipidemia (National Cholesterol Education Program, 2001). In populations where people are routinely physically active and lean, the atherogenic lipoprotein phenotype is minimally expressed. In sedentary populations that tend to be overweight or obese, very low fat, high carbohydrate diets clearly promote the development of this phenotype. Whether this phenotype promotes development of coronary atherosclerosis when it is specifically induced by low fat diets is uncertain, but it is a pattern that is associated with increased risk for CHD when expressed in the general American population. The atherogenic lipoprotein phenotype is characterized by higher triacylglycerol and decreased high density lipoprotein (HDL) cholesterol concentrations and small low density lipoprotein (LDL) particles. A predominance of small LDL particles is associated with a greater risk of CHD (Austin et al., 1990), but it is not known if this association is independent of increased triacylglycerol and decreased HDL cholesterol concentrations.

Table 11-2 and Figures 11-1 and 11-2 show that with decreasing fat and increasing carbohydrate intake, plasma triacylglycerol concentrations


TABLE 11-2 Fat and Carbohydrate Intake and Blood Lipid Concentrations in Healthy Individuals


Carbohydrate Intake (% of energy)

Coulston et al., 1983

11 men and women 10-d crossover P/S = 1.2-1.3

21 41

Bowman et al., 1988

19 men 10-wk parallel P/S = 0.4

29/60 33/58 45/42 46/42

Borkman et al., 1991

8 men and women 3-wk crossover

20/55 P/S = 0.46 50/31 P/S = 0.22

Kasim et al., 1993

72 women 1-y parallel P/S = 0.68-0.75

18 34

Leclerc et al., 1993

7 men and women 7-d crossover

11/64 30/45 40/45

Krauss and Dreon, 1995

105 men 6-wk crossover P/S = 0.69-0.74

24/60 46/39

O'Hanesian et al., 1996

10 men and women 10-d crossover

17/63 P/S = 0.25 28/57 P/S = 2.2 42/39 P/S = 1.7

Jeppesen et al., 1997

10 postmenopausal women 3-wk crossover P/S = 1.0

25/60 45/40

Kasim-Karakas et al., 1997

14 postmenopausal women 4-mo intervention

14 P/S = 1.2 23 P/S = 1.0 31 P/S = 0.9

Yost et al., 1998

25 men and women 15-d crossover P/S = 0.3

25/55 50/30

Straznicky et al., 1999

14 men 2-wk crossover

25/54 P/S = 1.3 47/36 P/S = 0.1

Kasim-Karakas et al., 2000

54 postmenopausal women 4- to 12-mo crossover P/S = 0.64

12/71 14/69 34/50


Postintervention Blood Lipid Concentration (mmol/L) b




1.51c 1.02d

0.98c 1.16d

0.91c 1.11c 0.84c 1.01c

1.42c 1.22c 1.53c 1.50c

2.35c 2.17c 2.59c 2.40c

0.82c (+49%) 0.55c

0.84c (-24%) 1.10d

2.88c (-20%) 3.60d

1.35c 1.25d

1.44c (-8%) 1.56d

2.79c (-10%) 3.09d

1.11c 1.29c 0.87d

1.03c 1.15d 1.32e

2.29c 2.47c 3.05d

1.59c 1.13d

1.09c 1.27d

<6 <¿<6

1.1 1.2 1.3

2.5 3.0

1.97c 1.29d

1.38c 1.49d

2.74c 2.81c

2.47c 2.10d 1.85e

1.24c 1.32d 1.34d

2.61c 2.93d 2.89d

1.14c 0.88d

1.22c 1.30d

0.8c 0.8c

1.05c 1.28d

2.6c 3.5d

1.49c 2.00c 1.57c

1.40c 1.29c 1.53d

3.49c 3.18c 3.57c



TABLE 11-2 Continued

Total Fat/

Carbohydrate Intake


Study Design^

(% of energy)


20 women

28/59 P/S = 0.7

et al., 2000

2-wk crossover

46/41 P/S = 0.4


459 men and

27/58 P/S = 1.1

et al., 2001b

women, 8-wk

37/52 P/S = 0.5


a P/S = polyunsaturated/saturated fatty acid ratio.

b HDL-C = high density lipoprotein cholesterol, LDL-C = low density lipoprotein cholesterol.

a P/S = polyunsaturated/saturated fatty acid ratio.

b HDL-C = high density lipoprotein cholesterol, LDL-C = low density lipoprotein cholesterol.

FIGURE 11-1 Relationship between percent of total fat intake and change in triacylglycerol (TAG) ( —) and high density lipoprotein (HDL) cholesterol (---) concentrations. Regression equations for percent change in serum TAG and HDL cholesterol predicted by percent total fat in the experimental diets of controlled-feeding studies comparing low fat, high carbohydrate diets to high fat diets. Weighted least-squares regression analyses were performed using the mixed procedure to test for differences in lipid concentrations (SAS Statistical package, version 8.00, SAS Institute, Inc., 1999). Percent of energy from total fat varied from 18.3% to 50%. All diets were low in saturated fat (less than 10% energy). Using these equations, for every 5% decrease in total fat, HDL cholesterol would decrease by 2.2% and triacylglycerol would increase by 6%.

DATA SOURCES: Berry et al. (1992); Curb et al. (2000); Garg et al. (1988, 1992a, 1994); Ginsberg et al. (1990); Grundy (1986); Grundy et al. (1988); Jansen et al. (1998); Kris-Etherton et al. (1999); Lefevre et al., unpublished; Lopez-Segura et al. (1996); Mensink and Katan (1987); Nelson et al. (1995); Parillo et al. (1992); Pelkman et al. (2001); Perez-Jimenez et al. (1995, 1999, 2001).


Postintervention Blood Lipid Concentration (mmol/L) b




0.81c 0.70d

1.34c 1.56d

2.43c 2.71d

c,d,e Within each study, LDL-C, HDL-C, or Lp(a) concentrations that are significantly different between treatment groups have a different superscript.

Ow 0


-jO ooc

h AO

50 60 70 80

Proportion of Energy Derived from Carbohydrates (%)

FIGURE 11-2 Relationship between proportion of energy from carbohydrates and serum high density lipoprotein (HDL) cholesterol concentration. • = Mean values for approximately 120 boys from five countries, o = individuals values for boys from the Philippines, FI= Finland, NE = Netherlands, GH = Ghana, IT = Italy, PH = Philippines.

SOURCE: Knuiman et al. (1987).


Dietary Total Fat (% energy)

FIGURE 11-3 Relationship between total fat intake and change in total cholesterol (TC):high density lipoprotein (HDL) cholesterol ratio. Weighted least-squares regression analyses were performed using the mixed procedure to test for differences in lipid concentrations (SAS Statistical package, version 8.00, SAS Institute, Inc., 1999).

DATA SOURCES: Berry et al. (1992); Curb et al. (2000); Garg et al. (1988, 1992a, 1994); Ginsberg et al. (1990); Grundy (1986); Grundy et al. (1988); Jansen et al. (1998); Kris-Etherton et al. (1999); Lefevre et al., unpublished; Lopez-Segura et al. (1996); Mensink and Katan (1987); Nelson et al. (1995); Parillo et al. (1992); Pelkman et al. (2001); Perez-Jimenez et al. (1995, 1999, 2001).

increase and plasma HDL cholesterol concentrations decrease. The reduction in HDL cholesterol concentration with low fat intake results in a rise in the total:HDL cholesterol concentration ratio (Figure 11-3). The total:HDL cholesterol ratio has been shown to be an important risk factor for CHD (Castelli et al., 1992; Kannel, 2000). Whether diet-induced changes in the total:HDL cholesterol ratio predispose to CHD remains unclear (Brussard et al., 1982; Jeppesen et al., 1997; Krauss and Dreon, 1995; West et al., 1990; Yost et al., 1998).

In support of the interventional studies, carbohydrate intake is negatively associated with HDL cholesterol concentrations (Table 11-3). Nonetheless, the association between atherogenic lipoprotein phenotype (higher

TABLE 11-3

Epidemiological Studies

on Carbohydrate Intake and Blood Lipid Concentrations


Study Design

Low Density Lipoprotein (LDL) Cholesterol Concentration

High Density Lipoprotein (HDL) Cholesterol Concentration

Triacylglycerol Concentration

Ernst et al., 1980

4,855 men and women Cross-sectional

Inversely related to carbohydrate intake

Knuiman et al., 1987

Multicountry regression analysis

Inversely related to carbohydrate intake

Fehily et al., 1988

653 men Cross-sectional regression analysis

No association

Negative association between carbohydrate intake and HDL concentration

No association

Multicountry regression analysis

Decreased with increased carbohydrate intake

Decreased with increased carbohydrate intake

Increased with increased carbohydrate intake

Tillotson et al., 1997

Prospective cohort,

6-y follow-up <29% carbohydrate 29-36% carbohydrate 36-41% carbohydrate 41-48% carbohydrate >48% carbohydrate


1.13 1.11 1.09 1.07 1.05

2.11 2.26 2.23 2.25 2.13


total:HDL cholesterol ratios) and CHD risk provides one rationale for establishing a lower boundary for the Acceptable Macronutrient Distribution Range (AMDR) for high-risk populations.

Risk of Hyperinsulinemia, Glucose Intolerance, and Type 2 Diabetes

Other potential abnormalities accompanying changes in distribution of fat and carbohydrate intakes include increased postprandial responses in plasma glucose and insulin concentrations. These abnormalities are more likely to occur with low fat, high carbohydrate diets. They potentially could be related to the development of both type 2 diabetes and CHD. In particular, repeated daily elevations in postprandial glucose and insulin concentrations could "exhaust" pancreatic P-cells of insulin supply, which could hasten the onset of type 2 diabetes. Some investigators have further suggested these repeated elevations could worsen baseline insulin sensitivity, which could cause susceptible persons to be at increased risk for type 2 diabetes. This form of diabetes, defined by an elevation of fasting serum glucose concentration, is characterized by two defects in glucose metabolism: insulin resistance, a defect in insulin-mediated uptake of glucose by cells, particularly skeletal muscle cells, and a decline in insulin secretory capacity by pancreatic P-cells (Turner and Clapham, 1998). Insulin resistance typically precedes the development of type 2 diabetes by many years. It is known to be the result of obesity, physical inactivity, and genetic factors (Turner and Clapham, 1998). Before the onset of diabetic hyperglycemia, the pancreatic P-cells are able to respond to insulin resistance with an increased insulin secretion, enough to maintain normoglycemia. However, in some persons who are insulin resistant, insulin secretory capacity declines and hyperglycemia ensues (Reaven, 1988, 1995).

The mechanisms for the decline in insulin secretion are not well understood, but one theory is that continuous overstimulation of insulin secretion by the presence of insulin resistance leads to "insulin exhaustion" and hence to decreased insulin secretory capacity (Turner and Clapham, 1998). Whether insulin exhaustion is secondary to a metabolic dysfunction of cellular production of insulin or to a loss of P-cells is uncertain. The accumulation of pancreatic islet-cell amyloidosis may be one mechanism for loss of insulin-secretory capacity (Hoppener et al., 2000).

High carbohydrate diets frequently causes greater insulin and plasma glucose responses than do low carbohydrate diets (Chen et al., 1988; Coulston et al., 1987). These excessive responses theoretically could predispose individuals to the development of type 2 diabetes because of prolonged overstimulation of insulin secretion (Grill and Bjorklund, 2001). The reasoning is similar to that for insulin resistance, namely, excessive stimulation of insulin secretion over a period of many years could result in


insulin exhaustion, and hence to hyperglycemia (Turner and Clapham, 1998). This mechanism, although plausible, remains hypothetical. Nonetheless, in the mind of some investigators, it deserves serious consideration.

Other consequences of hyperglycemic responses to high carbohydrate diets might be considered. For example, higher postprandial glucose responses might lead to other changes such as "desensitization" of P-cells for insulin secretion and production of glycated products or advanced glycation end-products, which could either promote atherogenesis or the "aging" process (Lopes-Virella and Virella, 1996). Again, these are hypothetical consequences that need further examination.

Epidemiological Evidence. A number of noninterventional, epidemiological studies have shown no relationship between carbohydrate intake and risk of diabetes (Colditz et al., 1992; Lundgren et al., 1989; Marshall et al., 1991; Meyer et al., 2000; Salmerón et al., 1997), whereas other studies have shown a positive association (Bennett et al., 1984; Feskens et al., 1991a).

Interventional Evidence. Interventional studies in healthy individuals on the influence of high carbohydrate diets on biomarker precursors for type 2 diabetes are lacking and the available data are mixed (Table 11-4) (BeckNielsen et al., 1980; Chen et al., 1988; Dunnigan et al., 1970; Fukagawa et al., 1990; Rath et al., 1974; Reiser et al., 1979). Factors such as carbohydrate quality, body weight, exercise, and genetics make the interpretation of such findings difficult. Nonetheless, in overweight and sedentary groups (which carry a heavy burden of insulin resistance and are common in North America), the accentuation of postprandial glucose and insulin concentrations that accompany high carbohydrate diets are factors to consider when setting an upper boundary for AMDRs for dietary carbohydrate (and a lower boundary for dietary fat).

Risk of Nutrient Inadequacy or Excess

Diets Low in Fats. For usual diets that are low in total fat, the intake of essential fatty acids, such as n-6 polyunsaturated fatty acids, will be low (Appendix K). In general, with increasing intakes of carbohydrate and decreasing intakes of fat, the intake of n-6 polyunsaturated fatty acids decreases. Furthermore, low intakes of fat are associated with low intakes of zinc and certain B vitamins.

The digestion and absorption of fat-soluble vitamins and provitamin A carotenoids are associated with fat absorption. Jayarajan and coworkers (1980) reported that the addition of 5 or 10 g of fat to a low fat (5 g) diet


TABLE 11-4 Intervention Studies on Carbohydrate Intake and Biochemical Indicators of Diabetes


Study Design

Dunnigan et al.,

9 men and women


4-wk crossover

31% sucrose


Rath et al.,

6 men


2- to 5-wk crossover

17% sucrose

52% sucrose

Reiser et al.,

19 men and women


6-wk crossover

30% starch

30% sucrose


7-d intervention

et al., 1980

Normal diet + 250 g glucose

Normal diet + 250 g fructose

Chen et al.,

8 men


3- to 5-d crossover

85% carbohydrate

41% carbohydrate

30% carbohydrate

Lundgren et al.,

1,462 women,


Prospective cohort,

12-y follow-up

Fukagawa et al.,

6 men


21- to 28-d intervention

40% carbohydrate

69% carbohydrate

a,b,c Within each study, the indicators of diabetes that are significantly different between treatment groups have a different superscript.



No diet effect on glucose tolerance and plasma insulin

Serum insulin (ng/mL) 5.4a 11.8b

Serum glucose (mg/dL)



Serum insulin (nmunits/mL) 9.8a 11.9b

Serum glucose (mg/dL)



No significant difference in insulin concentrations The high fructose diet was accompanied by a significant reduction in insulin binding and insulin sensitivity

Glucose disappearance Insulin sensitivity index (%/min)

Carbohydrate intake of women who developed diabetes (212 g/d) was not significantly different than women who did not develop diabetes (228 g/d)

Glucose disposal

Serum insulin (pmol/L) (pmol/kg/min)

67.4a 21.2a

50.2b 27.8b


significantly improved serum vitamin A concentrations. However, the addition of 10 g compared to 5 g did not provide any further benefit. The level of dietary fat has also been shown to improve vitamin K2 bioavailability (Uematsu et al., 1996). Dose-response data are limited on the amount of dietary fat needed to achieve the optimal absorption of fat-soluble vitamins, but it appears that the level is quite low.

Diets High in Fiber. Most diets that are high in fiber are also high in carbohydrate. High fiber diets have the potential for reduced energy density, reduced energy intake, and poor growth. However, poor growth is unlikely in the United States where most children consume adequate energy and fiber intake is relatively low (Williams and Bollella, 1995). Miles (1992) tested the effects of daily ingestion of 64 g or 34 g of Dietary Fiber for 10 weeks in healthy adult males. The ingestion of 64 g/d of Dietary Fiber resulted in a reduction in protein utilization from 89.4 to 83.7 percent and in fat utilization from 95.5 to 92.5 percent. Total energy utilization decreased from 94.3 to 91.4 percent. Because most individuals consuming high amounts of fiber would also be consuming high amounts of energy, the slight depression in energy utilization is not significant (Miles, 1992). In other studies, ingestion of high amounts of fruit, vegetable, and cereal fiber (48.3 to 85.6 g/d) also resulted in decreases in apparent digestibilities of energy, crude protein, and fat (Goranzon et al., 1983; Wisker et al., 1988). Again, however, the Dietary Fiber intakes were very high, and because the recommendation for Total Fiber intake is related to energy intake, the high fiber consumers would also be high energy consumers.

Diets High in Added Sugars. Increased consumption of added sugars can result in decreased intakes of certain micronutrients (Table 11-5). This can occur because of the abundance of added sugars in energy-dense, nutrient-poor foods, whereas naturally occurring sugars are primarily found in fruits, milk, and dairy products that also contain essential micro-nutrients. Because some micronutrients (e.g., vitamin B6, vitamin C, and folate), dietary fiber, and phytochemicals were not examined, the association between these nutrients and added sugars intakes is not known. Bowman (1999) used data from Continuing Survey of Food Intakes of Individuals (CSFII) (1994-1996) to assess the relationship between added sugars and intakes of essential nutrients in Americans' diets. The sample (n = 14,704) was divided into three groups based on the percentage of energy consumed from added sugars: (1) less than 10 percent of total energy (n = 5,058), (2) 10 to 18 percent of total energy (n = 4,488), and (3) greater than 18 percent of total energy (n = 5,158). Group 3, with a mean of 26.7 percent of energy from added sugars, had the lowest absolute mean intakes of all


the micronutrients, especially vitamin A, vitamin C, vitamin B-^, folate, calcium, phosphorus, magnesium, and iron. Compared with Groups 1 and 2, a decreased percentage of people in Group 3 met their Recommended Dietary Allowance (RDA) for many micronutrients. The individuals in Group 3 did not meet the 1989 RDA for vitamin E, vitamin B6, calcium, magnesium, and zinc. In addition, the high sugar consumers (Group 3) had lower intakes of grains, fruits, vegetables, meat, poultry, and fish compared with Groups 1 and 2. At the same time, Group 3 consumed more soft drinks, fruit drinks, punches, ades, cakes, cookies, grain-based pastries, milk desserts, and candies. Similar trends were also reported by Bolton-Smith and Woodward (1995) and Forshee and Storey (2001), but were not observed by Lewis and coworkers (1992). Emmett and Heaton (1995) reported an overall deterioration in the quality of the diet in heavy users of added sugars.

Using 1990-1991 cross-sectional data, Guthrie (1996) found that women whose diets met their RDA for calcium consumed significantly more milk products, fruit, and grains, and less regular soft drinks than women who did not meet their calcium recommendations. Others have shown that intakes of soft drinks are negatively related to intakes of milk (Guenther, 1986; Harnack et al., 1999; Skinner et al., 1999).

To further look at the association between added sugars and certain micronutrient intakes, the median intakes of various micronutrients at every 5th percentile of added sugars intake was determined using data from the Third National Health and Nutrition Examination Survey (NHANES III) (Appendix J). In addition, the prevalence of subpopulations not meeting the Estimated Average Requirement (EAR) or exceeding the Adequate Intake (AI) for these micronutrients was determined. Because not all micronutrients and other nutrients, such as fiber, were evaluated, it is not known what the association is between added sugars and these nutrients. While the trends are not consistent for all age groups, reduced intakes of calcium, vitamin A, iron, and zinc were observed with increasing intakes of added sugars, particularly at intake levels exceeding 25 percent of energy. Although this approach has limitations, it gives guidance for the planning of healthy diets.

Diets High in Total Sugars. In one large dietary survey, linear reductions were observed for certain micronutrients when total sugars intakes increased (Bolton-Smith and Woodward, 1995), whereas no consistent reductions were observed in another survey (Gibney et al., 1995) (Table 11-6). Bolton-Smith (1996) reviewed the literature on the relation of sugars intake to micronutrient adequacy and concluded that, provided consumption of sugars is not excessive (defined as less than 20 percent of total energy intake), no health risks are likely to ensue due to micronutrient inadequacies.


TABLE 11-5 Survey Data on Added Sugars and Micronutrient Intake


Diet Information

Nelson, 1991

143 children, 11-12 y

7-d weighed diet record

Rugg-Gunn et al., 1991

405 children, 11-14 y

3-d diet record

Lewis et al., 1992

Nationwide Food

Consumption Survey (1977-1978)

Bolton-Smith and Woodward, 1995

11,626 men and women,

25-64 y Scottish Heart Health and MONICA studies

Food frequency questionnaire

Gibson, 1997

1,675 boys and girls,

1.5-4.5 y U.K. National Diet and Nutrition Survey of Children

4-d weighed diet record

Bowman, 1999

Continuing Survey of Food Intakes by Individuals (1994-1996)

Two 24-h recalls

Forshee and Storey, 2001

Continuing Survey of Food Intakes by Individuals (1994-1996)


Added Sugars Intake (% of energy)

Change in Micronutrient Intake

16 2l 27

Decrease in nicotinic acid for girls l0 20

Decrease in vitamin D, protein

Percentile of intake

26th-75th > 75th

Decrease in calcium

Men: 1.0-6.2, 6.3-8.9, 9.0-13.0, 13.1-15.7, 15.8-47.9 Women: 0.8-4.8, 4.9-6.3, 6.4-8.1, 8.2-11.6, 11.7-50.2

Linear reduction in vitamin E, vitamin C, and vitamin A for both men and women

Decrease in zinc, calcium, riboflavin

Decrease in niacin, thiamin; large decrease in calcium, zinc, riboflavin

Decrease in calcium

Decrease in vitamin A, vitamin E, vitamin C, niacin, vitamin Bg, folate, vitamin B^, phosphorus, magnesium, iron, zinc, copper; large decrease in calcium

Negative correlation between added sugar intake and intake of vitamin A, calcium, and folate


TABLE 11-6 Survey Data on Total Sugar and Micronutrient Intake




Diet Information

Gibson, 1993

2,705 children Department of Health Survey of British School Children

7-d weighed food record

Bolton-Smith and Woodward, 1995

11,626 men and women,

25-64 y Scottish Heart Health and MONICA studies

Food frequency questionnaire

Gibney et al., 1995

Nicklas et al., 1996

8,296 men and women Nationwide Food

Consumption Survey (1987-1988)

568 children, 10 y Bogalusa Heart Study

3-d food record

24-h dietary recall

Farris et al., 1998 568 children, 10 y 24-h dietary recall

Bogalusa Heart Study

The impact of total sugar intake on the intake of micronutrients does not appear to be as great as for added sugars. Furthermore, a preliminary analysis of data from NHANES III on the intake of various micronutrients at every 5th percentile of total sugar intake did not reveal any significant associations as was observed for added sugars (Appendix J).

High Fat, Low Carbohydrate Diets of Adults Risk of Obesity

Epidemiological Evidence. Cross-country epidemiological data of dietary fat intake and obesity have yielded mixed results (Bray and Popkin, 1998;


Total Sugar Intake (% of energy)

Change in Micronutrient Intake

Decrease in iron, nicotinic acid Large decrease in iron, nicotinic acid No marked changes in micronutrient intake

Men: 2.5-12.0, 12.1-14.7, 14.8-17.2, 17.3-20.7, 20.8-51.4 Women: 1.5-11.7, 11.8-14.1, 14.2-16.3, 16.4-19.6, 19.7-52.8

Linear reduction in vitamin E, retinol, and vitamin A intake

Linear reduction in vitamin E, retinol, carotene, and vitamin A intake

Decrease in riboflavin, thiamin, calcium, iron, zinc, vitamin A Decrease in vitamin B6, vitamin E

Decrease in percent meeting the Recommended Dietary Allowance for niacin and zinc



Linear reduction in vitamin B6, vitamin E, thiamin, iron, zinc, and niacin intake with increasing total sugar intake

Willett, 1998). In some countries, low fat, high carbohydrate diets are associated with a low prevalence of obesity, whereas in others they are not.

Within-country surveys of dietary intake and body mass index (BMI) have also yielded mixed results. Many case-control and prospective studies failed to find a strong correlation between percent of energy intake from fat and body weight (Heitmann et al., 1995; Lissner et al., 2000; Ludwig et al., 1999b; Rissanen et al., 1991; Samaras et al., 1998; Willett, 1998), whereas some did find significant associations (Bray and Popkin, 1998; Dreon et al., 1988; George et al., 1990; Klesges et al., 1992; Miller et al., 1990; Romieu et al., 1988; Tucker and Kano, 1992). Colditz and coworkers (1990) observed no association between fat intake and weight gain prospectively, but did find a positive association between previous weight


gain and high fat intake. One statistically well-designed study that included direct measurements of body fat and considered potentially confounding factors such as exercise concluded that total dietary fat was positively correlated with fat mass (adjusted for fat-free mass, r = 0.22, p < 0.0001) in adults (Larson et al., 1996). Most multiple regression studies found that about 3 percent of the total variance in body fatness was explained by diet, though some studies placed the estimate at 7 to 8 percent (Westerterp et al., 1996). Longitudinal studies generally supported dietary fat as a predictive factor in the development of obesity (Lissner and Heitmann, 1995). However, bias in subject participation, retention, and underreporting of intake may limit the power of these epidemiological studies to assess the relationship between dietary fat and obesity or weight gain (Lissner et al., 2000).

Another line of evidence often cited to indicate that dietary fat is not an important contributor to obesity is that although there has been a reduction in the percent of energy from fat consumed in the United States, there has been an increase in energy intake and a marked gain in average weight (Willett, 1998). Survey data showed an increase in total energy intake over this period (McDowell et al., 1994), so that despite the decline in percent of energy from fat, the total intake of fat (g/d) remained stable. Another study that used food supply data showed that fat intake may indeed be rising in the United States (Harnack et al., 2000).

Mechanisms for Obesity and Interventional Evidence. Several mechanisms have been proposed whereby high fat intakes could lead to excess body accumulation of fat. Foods containing high amounts of fat tend to be energy dense, and the fat is a major contributor to the excess energy consumed by persons who are overweight or obese (Prentice, 2001). The energy density of a food can be defined as the amount of metabolizable energy per unit weight or volume (Yao and Roberts, 2001); water and fat are the main determinants of dietary energy density. Energy density is an issue of interest to the extent that it influences energy intake and thus plays a role in energy regulation, weight maintenance, and the subsequent development of obesity.

Three theoretical mechanisms have been identified by which dietary energy density may affect total energy intake and hence energy regulation (Yao and Roberts, 2001). Some studies suggest that, at least in the short-term, individuals tend to eat in order to maintain a constant volume of food intake because stomach distension triggers vagal signals of fullness (Duncan et al., 1983; Lissner et al., 1987; Seagle et al., 1997; Stubbs et al., 1995a). Thus, consumption of high energy-dense foods could lead to excess energy intake due to the high energy density to small food volume ratio.


A second proposed mechanism is that high energy-dense foods are often more palatable than low energy-dense foods (Drewnowski, 1999; Drewnowski and Greenwood, 1983). A survey of American adults reported that taste is the primary influence for food choice (Glanz et al., 1998). In single-meal studies, high palatability was also associated with increased food consumption (Bobroff and Kissileff, 1986; Price and Grinker, 1973; Yeomans et al., 1997). These results suggest that high energy-dense foods may be overeaten because of effects related to their high palatability.

The third mechanism is that energy-dense foods reduce the rate of gastric emptying (Calbet and MacLean, 1997; Wisen et al., 1993). This reduction, however, does not occur proportionally to the increase in energy density. Although energy-dense foods reduce the rate at which food leaves the stomach, they actually increase the rate at which energy leaves the stomach. Thus, because energy-containing nutrients are digested more quickly, nutrient levels in the blood fall quicker and hunger returns (Friedman, 1995). While a subjective measure, highly palatable meals have also been shown to produce an increased glycemic response compared with less palatable meals that contain the same food items that are combined in different ways (Sawaya et al., 2001). This suggests a generalized link among palatability, gastric emptying, and glycemic response in the underlying mechanisms determining the effects of energy density on energy regulation. Further research on this potential link is needed.

Researchers have used instruments such as visual analogue scales to measure differences in appetite sensations (e.g., hunger and satiety) between treatments in order to examine the effects of altering nutrients that play a major role in energy density, such as dietary fat, on energy regulation (Flint et al., 2000). A number of studies have been conducted in which preloads of differing energy density were given and hunger and satiety were measured either at the subsequent meal or for the remainder of the day. In the studies that administered preloads that had constant volume but different energy content (energy density was altered by changing dietary fat content), there was no consistent difference in subsequent satiety or hunger between the various test meals (Durrant and Royston, 1979; Green et al., 1994; Hill et al., 1987; Himaya et al., 1997; Hulshof et al., 1993; Louis-Sylvestre et al., 1994; Porrini et al., 1995; Rolls et al., 1994). However, in those studies using isoenergetic preloads that differed in volume (energy density was altered by changing dietary fat content), there was consistently increased satiety and reduced hunger after consumption of the low energy-dense preload meals (i.e., those with higher volume) (Blundell et al., 1993; Holt et al., 1995; van Amelsvoort et al., 1989, 1990). It has been reported, however, that diets low in fat and high in carbohydrate may lead to more rapid return of hunger and increased snacking between meals (Ludwig et al., 1999a).


These data suggest that in the short-term, low energy-dense foods appear to increase satiety and decrease hunger compared to high energy-dense foods. Because individuals were blinded to the dietary content of the treatment diets, the results from these studies demonstrate the short-term effects of energy density after controlling for cognitive influences on food intake.

It is important that cognitive factors are taken into account during the interpretation of results of preload studies. When individuals were aware of dietary changes, they generally (Ogden and Wardle, 1990; Shide and Rolls, 1995; Wooley, 1972), but not always (Mattes, 1990; Rolls et al., 1989), compensated for changes in energy density and thus minimized changes in energy intake.

In well-controlled, short-term intervention studies lasting several days or more, high fat diets were consistently associated with higher spontaneous energy intake (Lawton et al., 1993; Proserpi et al., 1997; Thomas et al., 1992). From short- and longer-term studies, volunteers consistently consumed less dietary energy on low fat, low energy dense diets compared to high energy-dense diets (Glueck et al., 1982; Lissner et al., 1987; Poppitt and Swann, 1998; Poppitt et al., 1998; Stubbs et al., 1995b; Thomas et al., 1992; Tremblay et al., 1989, 1991). The extent to which energy intake was reduced on low energy-dense diets was similar for short- and long-term studies.

An alternative way to study the effects of energy density on energy intake in short-term studies has been to compare energy intake between diets of similar energy density that differ in dietary fat content. Using this approach, when fat content was covertly varied between 20 and 60 percent of energy, there was no significant difference in energy intake between groups (Saltzman et al., 1997; Stubbs et al., 1996; van Stratum et al., 1978). These results suggest that energy density plays a more significant role than fat per se in the short-term regulation of food intake.

During overfeeding, fat may be slightly more efficiently used than carbohydrate (Horton et al., 1995), but in one study, no difference was seen (McDevitt et al., 2000). Thus, high fat diets are not intrinsically fattening, calorie for calorie, and will not lead to obesity unless excess total energy is consumed. It is apparent, however, that with the consumption of high fat diets by the free-living population, energy intake does increase, therefore predisposing to increased weight gain and obesity if activity level is not adjusted accordingly (see Table 11-1). While many of the short-term studies showed a more dramatic effect on weight reduction with reduced fat intake, the long-term studies showed weight loss as well.

Conclusions. Epidemiological studies provide mixed results on the question of whether high fat (low carbohydrate) diets predispose to over-


weight and obesity and promote weight gain. However, a number of short-term studies suggest mechanisms whereby high fat intake could promote weight gain in the long-term. In addition, short- and long-term intervention studies provide evidence that reduced fat intake is accompanied by reduced energy intake and therefore moderate weight reduction or prevention of weight gain. For these reasons, it may be concluded that higher fat intakes are accompanied with increased energy intake and therefore increased risk for weight gain in populations that are already disposed to overweight and obesity, such as that of North America.

Lose 10 Pounds Naturally

Lose 10 Pounds Naturally

Studies show obesity may soon overtake tobacco as the leading cause of death in America. Are you ready to drop those extra pounds you've been carrying around? Awesome. Let's start off with a couple positive don't. You don't need to jump on a diet craze and you don't need to start exercising for hours each day.

Get My Free Ebook

Post a comment