点击显示 收起
1 From the General Clinical Research Center (JP and RH), Department of Medicine and Cardiovascular Center (JJ and DG), and Department of Pediatrics (AS), Medical College of Wisconsin, Milwaukee
2 Supported by the General Clinical Research Center grant M01-RR00058 to the Medical College of Wisconsin from the National Institutes of Health. 3 Reprints not available. Address correspondence to A Slyper, Department of Pediatrics, Loyola University Medical Center, 2160 South First Avenue, Maywood, IL 60153. E-mail: aslyper{at}lumc.edu.
ABSTRACT
Background: The influence of dietary carbohydrate glycemic index on blood lipids has not been well studied. Assessment of glycemic load is not usually included in a standard dietary analysis.
Objective: The purpose of the present study was to examine relations between diet and blood lipids in youth with a broad range of cholesterol values and carbohydrate, fat, and protein intakes.
Design: Relations between blood lipids and dietary constituents were examined in 32 healthy males and females aged 11–25 y. Subjects exhibited a range of LDL-cholesterol values (1.71–6.67 mmol/L) and body mass index z scores (–1.18 to 2.64). Dietary constituents were assessed from 3-d food diaries.
Results: The only significant correlations evident were negative correlations between HDL cholesterol and glycemic load (in relation to white bread), percentage carbohydrate, total dietary sugar, total carbohydrate, and fructose. On stepwise multiple regression analysis, glycemic load accounted for 21.1% of the variation in HDL cholesterol.
Conclusions: Glycemic load appears to be an important independent predictor of HDL cholesterol in youth. This relation is of concern and suggests that dietary restrictions without attention to a possible resulting increase in glycemic load may result in an unfavorable influence on blood lipids.
Key Words: Dietary carbohydrate • glycemic load • HDL cholesterol
INTRODUCTION
Traditionally, relations between diet and blood lipids were examined in terms of the main food groups: fat, protein, and carbohydrate. Dietary carbohydrate is usually further divided into simple sugars and complex carbohydrate. In terms of their postprandial metabolic and hormonal responses, however, many complex carbohydrates are little different from simple sugars, and it makes considerable sense, therefore, to classify carbohydrates in terms of their postprandial glucose responses. This can be conveniently expressed in terms of "glycemic index" with respect to either glucose or white bread (1). Glycemic load reflects both the glycemic index of dietary carbohydrate as well as the amount of carbohydrate ingested (2). The purpose of this study was to examine for relations between diet and conventional lipids factors in a group of youth who displayed a wide range of cholesterol values and hence a broad range of fat, protein, and carbohydrate intake.
SUBJECTS AND METHODS
Subjects were healthy males and females aged 11–25 y. Some of the subjects had attended the Pediatric Lipid Clinic at the Children's Hospital of Wisconsin because of hyperlipidemia or a family history of coronary disease. Other subjects had responded to advertisements inviting their participation without respect to blood lipids or family history of coronary disease. Because of this recruitment strategy, participants exhibited a broad range of blood lipids and body mass index z scores. The study protocol was approved by the Institutional Review Board of the Children's Hospital of Wisconsin.
Subjects attended the Clinical Research Center on the day of the study in the fasting state after a 12-h fast. After signing informed consent, subjects underwent a brief physical examination, including height, weight, and blood pressure. Blood was drawn into EDTA-coated tubes for blood lipids. Subject characteristics and their lipid values are shown in Table 1.
View this table:
TABLE 1. Characteristics and laboratory values of the 32 subjects1
Study participants, and their parents when appropriate, were instructed by the research dietitian on how to record food intake for 3 d (2 weekdays and 1 weekend day). The completed food diaries were reviewed with the subject by the same dietitian who used food models to document quantities. Nutrient calculations were performed with the use of the Nutrition Data System for Research software version 4.05 developed by the Nutrition Coordinating Center, University of Minnesota, Minneapolis, and Food Nutrient Database 33, released July 2002 (1). Mean dietary constituents were estimated as grams per day and percentage carbohydrate, fat, and protein as the percentage of daily calories. The glycemic index of each food was multiplied by the carbohydrate content of that food, and the sum of those values provided the average daily glycemic load. White bread was used as the standard for calculating glycemic load. Foods that did not have a glycemic value were assigned a value by using a similar food. A program was written to integrate the values for glycemic index and glycemic load into the Nutrition Data System for Research software (3). Milk fat and milk protein from milk alone were also estimated, as well as dairy fat and dairy protein from the dairy products milk, cheese, yogurt, and ice cream. Dietary intakes of the subjects are summarized in Table 2.
View this table:
TABLE 2. Daily dietary intakes of the 32 subjects
Blood lipids were determined at the Emory Lipid Research Laboratory, which is a participant in the Lipid Standardization Program of the Centers for Disease Control and Prevention and the National Heart, Lung, and Blood Institute. All analyses were performed on freshly isolated fasting plasma from EDTA-coated tubes with use of a Beckman CX7 chemistry autoanalyzer (Beckman Coulter Diagnostics, Fullerton, CA). Total triacylglycerol and cholesterol were determined by enzymatic methods (Beckman Coulter Diagnostics). Direct HDL and direct LDL cholesterol was obtained with use of homogeneous assays (Equal Diagnostics, Exton, PA). Apolipoprotein A-I and apoprotein B were determined with use of immunoturbidometric reagents (DiaSorin, Stillwater, MN). Triacylglycerol-rich lipoproteins were isolated by ultracentrifugation at density d < 1.020 g/mL, thereby including VLDL and intermediate-density lipoprotein. The supernate was quantitatively recovered in volumetric tubes, and its content of triacylglycerol and cholesterol was determined by enzymatic methods as described above as a measure of lipoprotein remnants.
Correlations and multiple regression analysis were determined with use of the MINITAB statistical package (State College, PA).
RESULTS
The only significant correlations between dietary constituents and blood lipids were with HDL cholesterol. Significant correlations with HDL cholesterol were evident for glycemic load, glycemic index, total sugars, total carbohydrate, and fructose (Table 3). The correlation between glycemic load and HDL cholesterol is demonstrated in Figure 1. Stepwise multiple regression analysis was performed with all the above variables being entered into the regression analysis, and glycemic load was the only independent predictor of HDL cholesterol, accounting for 21.1% of its variation.
View this table:
TABLE 3. Significant dietary correlates with HDL cholesterol
FIGURE 1.. Correlation between glycemic load and HDL cholesterol.
Significant correlations between daily glycemic load and daily dietary constituents are shown in Table 4. Of note are the negative correlations between glycemic load and percentage protein and fat and the positive correlations between glycemic load and protein and fat intake. These correlations are displayed in Figure 2 A to D.
View this table:
TABLE 4. Significant dietary correlates with glycemic load
FIGURE 2.. Correlations between glycemic load and fat, glycemic load and percentage fat, glycemic load and protein, and glycemic load and percentage protein.
DISCUSSION
Interpretation of data from this study indicates that glycemic load is an important independent predictor of HDL cholesterol, accounting for more than 20% of its variation. This conclusion is in accord with that of Liu et al (4), who showed that the quality (as measured by glycemic index) and quantity of carbohydrate were directly related to HDL cholesterol and triacylglycerol in 280 healthy postmenopausal women and that glycemic load best captured the limits of carbohydrate intake. The Third National Health and Nutrition Examination Survey (1988–1994) also demonstrated among their 13 000 participants a relation between HDL cholesterol and glycemic index and glycemic load (5).
In those 2 previous studies, glycemic load was estimated by means of food-frequency questionnaires, whereas the present study used 3-d food records. The latter provides a more precise indication of the kinds and quantities of food eaten than a food-frequency questionnaire. When using food-frequency questionnaires, subjects are asked to recall what they have eaten in the past, whereas with food records they are asked to record what they have consumed over a specific time period. Furthermore, food-frequency questionnaires usually use a defined list of foods, and this does not permit a precise measure of the quantity of food eaten. Concerns with food records are that subjects may alter their dietary intake during the period of assessment, forget to record items, or be imprecise in measuring amounts eaten. Nevertheless, a thorough review of the food record by a trained nutritionist can reduce these errors. A shortcoming of both methodologies is the lack of published glycemic index values for many food items. In this present study, estimates were made for unknown foods by using similar foods of known value.
This particular group of subjects was studied because it contained individuals who had a wide range of cholesterol and body weights. This group would, therefore, demonstrate a broader range of dietary practices than would be found in a similar-sized group of healthy subjects.
Findings from the Bogalusa Heart Study showed that children occupying the lowest percentiles of fat consumption increase their intake of simple sugars (6). Gibney (7) also pointed out, on the basis of cross-sectional data from several European countries, that low-fat diets are invariably associated with an increase in the consumption of simple sugars. In the United States, consumption of milk declined by 36% between 1965 and 1996, and the lost calories were replaced mainly by soft drinks, soda, and fruit drinks (8, 9).
It is logical to assume that when individuals reduce fat and protein, either on their own volition or as a result of dietary counseling, these calories will be replaced by dietary carbohydrate. Many of the carbohydrate items readily available to American teenagers and young adults, such as ready-to-eat cereals, potatoes, white bread, and snack foods, have a high glycemic index. In fact, interpretation of data from this study does show moderate negative correlations between glycemic load and percentage dietary protein and fat. This study highlights the concern that lipid-lowering diets may inadvertently lead to an increase in glycemic load and thereby a reduction in HDL cholesterol.
This study did demonstrate a positive relation between glycemic load and dietary protein and fat intake. This relation probably reflects the fact that, as total carbohydrate and calories increase, so does the amount of dietary protein and fat. However, the negative correlations between glycemic load and percentage dietary fat and protein probably reflect that, as dietary fat and protein decrease, there is an increase in the amount of high glycemic carbohydrate.
Interpretation of data from this study highlights a possible concern with respect to dissecting dietary influences on cardiovascular disease. In this study, glycemic load was moderately correlated with many other dietary constituents, including total, saturated, monounsatured, and polyunsatured fats; total protein; and milk protein. Glycemic load was not a component of most previous dietary studies, and this raises the possibility that a true relation between glycemic load and various aspects of cardiovascular disease was incorrectly ascribed to other dietary factors. In a large 10-y prospective study in women aged 38–63 y, Liu et al (10) noted that dietary glycemic load was directly associated with risk of coronary artery disease after adjustment for smoking, age, total energy intake, and other coronary risk factors. This relation was also present after adjusting for dietary fat, including saturated and trans fats.
In conclusion, this study demonstrates an inverse relation between glycemic load and HDL cholesterol in a group of hyperlipidemic and normolipidemic older children, teenagers, and young adults. HDL cholesterol is well recognized as an important independent coronary risk factor in healthy individuals and those with coronary artery disease (11–13). A causal relation in this instance cannot be excluded. It is suggested that lipid-lowering diets have the potential for inducing unfavorable consequences on blood lipids. Counseling on lipid-lowering diets should be provided only by trained professionals, and advice about appropriate carbohydrate needs should be an integral component of this counseling.
ACKNOWLEDGMENTS
AS, JJ, and DG were involved in the design and performance of this study. RH was involved in the statistical design and analysis of the study. JP was involved in the performance of this study and the dietary analysis. None of the authors had any conflicts of interest with respect to this study.
REFERENCES