| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Department of Human Nutrition, Centre for Advanced Food Studies, The Royal Veterinary and Agricultural University, Frederiksberg, Denmark (BS, IK-M, AF, IT, AA and AR); Danone Vitapole, Paris (SV and VL); and the Department of Applied Nutrition & Food Chemistry, Lund University, Sweden (IB and HE)
2 Supported by Danone Vitapole, France. Rice was donated by Masterfoods a.s., Denmark, and Euryza GmbH, Germany, and rye bread was donated by Cerealia R&D, Schulstad Brød A/S, Denmark. 3 Reprints not available. Address correspondence to B Sloth, Department of Human Nutrition, Centre for Advanced Food Studies, The Royal Veterinary and Agricultural University, 30 Rolighedsvej, DK-1958 Frederiksberg C, Denmark. E-mail: bsl{at}kvl.dk.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: The objective of the study was to investigate the long-term effects of a low-fat, high-carbohydrate diet with either low glycemic index (LGI) or high glycemic index (HGI) on ad libitum energy intake, body weight, and composition, as well as on risk factors for type 2 diabetes and ischemic heart disease in overweight healthy subjects.
Design: The study was a 10-wk parallel, randomized, intervention trial with 2 matched groups. The LGI or HGI test foods, given as replacements for the subjects’ usual carbohydrate-rich foods, were equal in total energy, energy density, dietary fiber, and macronutrient composition. Subjects were 45 (LGI diet: n = 23; HGI diet: n = 22) healthy overweight [body mass index (in kg/m2): 27.6 ± 0.2] women aged 20–40 y.
Results: Energy intake, mean (± SEM) body weight (LGI diet: –1.9 ± 0.5 kg; HGI diet: –1.3 ± 0.3 kg), and fat mass (LGI diet: –1.0 ± 0.4 kg; HGI diet: –0.4 ± 0.3 kg) decreased over time, but the differences between groups were not significant. No significant differences were observed between groups in fasting serum insulin, homeostasis model assessment for relative insulin resistance, homeostasis model assessment for ß cell function, triacylglycerol, nonesterified fatty acids, or HDL cholesterol. However, a 10% decrease in LDL cholesterol (P < 0.05) and a tendency to a larger decrease in total cholesterol (P = 0.06) were observed with consumption of the LGI diet as compared with the HGI diet.
Conclusions: This study does not support the contention that low-fat LGI diets are more beneficial than HGI diets with regard to appetite or body-weight regulation as evaluated over 10 wk. However, it confirms previous findings of a beneficial effect of LGI diets on risk factors for ischemic heart disease.
Key Words: Obesity • fat mass • energy intake • type 2 diabetes • ischemic heart disease • cholesterol • triacylglycerol • glucose • insulin
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Several single-meal studies of appetite and GI have been performed, but with inconsistent results (5–7). Furthermore, those studies have not yet been followed by enough longer-term studies to allow conclusions as to the long-term effects of low-GI (LGI) versus high-GI (HGI) foods on voluntary energy intake and body weight. Only 1 longer-term study (5 wk) with diets of different GIs has been designed as an ad libitum study (8), whereas other studies have been designed as weight-maintaining or energy-restricted studies (6). In general, the diets in the previous studies were not well-matched. Besides the GI, macronutrient composition, dietary fiber content, energy density, or any combination of those factors also differed between test diets. To study the effect of GI per se, all other dietary factors should be kept constant.
Better glycemic control with LGI diets than with HGI diets has been shown in subjects with type 2 diabetes (9) and in subjects with impaired glucose tolerance (10). However, the relevance of GI as a mean of preventing insulin resistance and type 2 diabetes in healthy nondiabetic persons is still controversial, and confounding factors such as cereal fiber and whole-grain consumption complicate the picture (11). In addition, several epidemiologic studies have given evidence to support connections between a high dietary GI and ischemic heart disease (IHD) or risk markers for IHD (12–14). Most intervention studies were, however, performed in diabetic subjects, and therefore little evidence exists for the relevance of GI to IHD risk factors in healthy persons.
The purpose of this study was to investigate the effect of a low-fat, high-carbohydrate diet with either LGI or HGI carbohydrates—all other dietary components being equal—on ad libitum energy intake, body weight, and body composition and on risk markers for type 2 diabetes and IHD after a 10-wk intervention in healthy, slightly overweight subjects.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
On the first day and the last day (wk 10) of the study, we measured body weight, height, blood pressure, heart rate, sagittal height, waist and hip circumference, and body composition (by using dual-energy X-ray absorptiometric scanning) and collected fasting blood samples. Body weight was measured in weeks 2, 4, 6, and 8. Subjects completed a 7-d weighed dietary record just before entering the study and in weeks 5 and 10 of the study period. During each of these periods, a 24-h urine collection was made to determine urinary nitrogen excretion.
Subjects
Power calculations made before the study showed that a total of 43 subjects was needed to obtain a significant (P < 0.05) difference in body weight change of 2.0 kg (SD = 2.0) with a power of 90%. Women were recruited from the area of Copenhagen and Frederiksberg by local newspaper advertisement and from local universities by posted announcements.
The inclusion criteria for the study included age 20–40 y, body mass index (in kg/m2) 25–30, body weight fluctuations 
5 kg over the previous 2 mo, absence of any physiologic or psychological illnesses that could influence the study results, no regular use of medicine other than birth control pills, normal to mildly hypertensive blood pressure (159/99 mm Hg; 15), no allergies to any foods, no special diets (eg, vegetarian) or particular dislikes, consumption of 14 alcoholic drinks/wk (1 drink equaling 10 g alcohol), and nonsmoker status (defined as 1 cigarette/d). Subjects were also required not to be elite athletes or to be planning to change physical activity during the study; to be nonpregnant and with no pregnancy planned within the study period, nonlactating, and premenopausal with regular menstrual cycle; and to have made no blood donation within the past 3 mo before entering the study.
Approximately 500 persons responded by telephone or e-mail to the recruitment effects; 300 of these persons were sent written information. After reading this information, 153 of the 300 persons contacted us to book a screening test for measurement of body weight, height, and blood pressure and an interview regarding general health and drinking and smoking habits. A total of 55 suitable and willing subjects were enrolled in the study.
The subjects were randomly assigned to the 2 intervention groups, which were matched for age, body weight, height, body mass index, blood pressure, heart rate, estimated energy expenditure (16), and alcohol intake. Of the 55 enrolled subjects, 48 completed the study, which resulted in a 13% dropout rate. Two subjects in the HGI diet group dropped out, one after 1 wk because she found the study too demanding and one after 5 wk for personal reasons. Five subjects in the LGI diet group dropped out: one was living with one of the dropouts from the HGI diet group and also dropped out after 1 wk because she found the study too demanding; 2 subjects gave the same reason for dropping out after 2 wk and also mentioned that the amount of test food was too much; 1 subject dropped out because of a staphylococcus infection; and 1 subject dropped out because her son was hospitalized. Of the 48 subjects who completed the study, 2 subjects in the HGI diet group did not comply fully with the protocol, as determined from evaluation of their self-reported data on amounts of the test foods eaten, and their data were excluded from analysis. One subject in the LGI diet group discovered that she was pregnant just after she completed the study, and her data were also excluded from all analyses.
Characteristics of the 45 subjects whose data were analyzed are presented in Table 1
. There were no significant differences between the groups in baseline values. Subjects’ maximum body weight during their lifetime (apart from pregnancy periods), given as the percentage of difference from the current weight, ranged from 0.0% to 17.3% (
± SEM: 5.1 ± 0.7%), and there was no significant difference between groups.
|
30% of energy from fat) diets that were rich in either LGI or HGI foods. To ensure that subjects did indeed consume a HGI or LGI diet, they received a certain amount of carbohydrate-rich test foods from the Department of Human Nutrition every week. The test foods consisted of wheat bread, rye bread (the bread eaten daily by most Danes), rice and pasta or mashed potato powder. In the LGI diet group, test foods were whole-grain wheat bread (grains from Valsemøllen A/S, Esbjerg, Denmark; baked at the Department of Human Nutrition), whole-grain rye bread (Schulstad Brød A/S, Copenhagen), long-grain rice (Uncle Ben’s; Masterfoods, Olen, Belgium), and pasta (Al Dente; Irma, Rødovre, Denmark). By whole grain, we mean intact whole kernels. In the HGI diet group, test foods were whole-meal wheat bread (flour from Valsemøllen A/S; baked at the Department of Human Nutrition), whole-meal rye bread (Schulstad Brød A/S), round grain rice (maximum 4% of grains broken, from Euryza, Hamburg, Germany), and mashed potato powder (Eurogran, Kalundborg, Denmark). The whole-meal flour was made from kernels similar to those used in the LGI diets. Macronutrient composition was kept similar in the 2 groups of test foods by adjusting protein and fat intake with a low-fat sour milk product (0.3% fat; Fromage Frais; MD Foods, Aarhus, Denmark) and butter (80% fat; Kærgården; MD Foods). In addition, energy density was kept similar by adding water to the menus (Table 2).
|
). They were instructed to measure and record the exact amounts eaten every day by using a digital scale and a food diary. Energy requirements were calculated from body weight, height, age, and physical activity according to the questionnaire of Baecke et al (17) by using the FAO/UNU/WHO formulas (16), and subjects were accordingly categorized as having 8, 10, or 12 MJ/d energy requirements and thus were assigned to test food intakes of 3.9, 4.9, or 5.8 MJ/d. The amount of test foods given should account for 
75% of total carbohydrate intake, on the assumption that 55% of ingested energy would be from carbohydrates. Subjects were also instructed to eat a diet with 20–30% of energy from fat, and a list with other carbohydrate-rich foods was given to the subjects so that they could monitor the GI of any such foods from their usual diets that they ate during the study. This list also instructed subjects to have a low sugar intake, because we wanted the carbohydrates eaten to derive primarily from starchy foods. Subjects were to limit cake and candy intake to a minimum; consume only artificially sweetened soft drinks, jams, and yogurts; and limit juice and chocolate milk intake to a maximum of 0.5 L/d. Apart from the test foods, subjects could eat ad libitum of their own diet until pleasantly satisfied. The subjects received individual guidance by trained clinical dietitians on the first day of the study period and at group meetings in weeks 3, 5, 7, and 9.
|
All nutrient calculations of test foods and 7-d dietary records were done with the use of a computer database of foods from The National Food Agency of Denmark (Dankost 2000) that included specific product-ingredient lists and recipes for test foods. To identify underreporting, energy intake/basal metabolic rate (EI/BMR) was calculated on the basis of the 7-d dietary record periods for each subject for comparison with Goldberg’s cutoff limits (18). BMR was calculated according to Schofield’s equation (19) by using baseline body weight for baseline EI/BMR, mean body weight from weeks 4 and 6 for week 5 EI/BMR, and week 10 body weight for week 10 EI/BMR.
Carbohydrate quality of test foods
Carbohydrate quality of the test foods was determined with the use of 2 in vitro methods, both of which are well correlated with GI measured in vivo (20–24). First, a hydrolysis index of the breads, pasta, rice, and mashed potato was determined at the Department of Applied Nutrition and Food Chemistry (Lund University, Sweden) by using white bread as a reference (20). This in vitro method is based on enzymatic incubation of chewed samples that are equivalent to 1 g starch. The available starch in food samples was determined by enzymatic analysis (25). Relation between hydrolysis index (HI) and GI values is given by the formula GI = 0.862 HI + 8.198 (26). The weighted average GIs of the test foods were 78.6 and 102.8 for the LGI and HGI foods, respectively, which resulted in a difference between groups of 24.3 units (Table 2
). Weighted GI was calculated by the method given by FAO/WHO (27).
Second, the breads, pasta, rice, and mashed potato were analyzed by the method of Englyst et al (28–30) for total carbohydrate, total fructose, total glucose, total starch, resistant starch, rapidly available glucose, and slowly available glucose (Table 3). Total starch, total glucose, and total carbohydrate content were higher in the LGI products, and this difference is at least partly explained by the difference in dry matter. In particular, the mashed potatoes in the HGI diet had a higher water content than did the pasta, which was meant to be an LGI match for the potatoes. There was almost no fructose in the 8 test foods and only small quantities of resistant starch, and there were no differences between groups.
|
Measurements
All measurements were performed in the morning after subjects had fasted for 
10 h. Height was measured to the nearest 0.5 cm in subjects not wearing shoes by using a wall-mounted stadiometer. The body weight of subjects dressed in underwear was measured to the nearest 0.05 kg with the use of a digital scale (Lindeltronic 8000, D93-09–141; Lindell, Lindells, Sweden) after the subject had voided. Sagittal height (height of abdomen after exhalation while the subject lay in a supine position) was measured to the nearest 0.5 cm. Waist circumference was measured to the nearest 0.5 cm at the narrowest point between the iliac crest and the lowest rib. Hip circumference was measured to the nearest 0.5 cm at the broadest point below the iliac crest. Body composition was assessed with the use of whole-body dual-energy X-ray absorptiometry scanning by using a Lunar DPX-IQ scanner (General Electrics, Madison, WI) while the subjects were dressed in light clothing and not wearing any metal (31, 32). Blood pressure and heart rate were measured with the use of a fully automatic blood pressure monitor (Omron M4-I; Omron Healthcare Europe BV, Hoofdorp, Netherlands). A mean of 2 measurements was used. Subjects rested in a supine position with the head slightly elevated for 15 min before the blood pressure measurements.
A 24-h urine sample was collected on day 3 (of weeks 0 and 10) or day 5 (of week 5) of the three 7-d dietary record periods. As a marker of complete urine samples (33), 3 tablets with a total of 240 mg 4-aminobenzoic acid (PABA) were taken at meal times. The volume and density of each 24-h urine collection were determined, and a sample was frozen at –20 °C. Urine samples were analyzed for aromatic amines (PABA) by colorimetric method using a spectrophotometer (Stasar; Gilford Instruments Laboratories Inc, Oberlin, OH) with intraassay and interassay CVs <5% and <7%, respectively. Urine samples with a PABA recovery <85% were considered incomplete urine collections.
To calculate nitrogen recovery and validate protein intake (34) from dietary records, 24-h urinary nitrogen excretion was determined with the use of the Dumas method (35) and a nitrogen analyzer (NA1500; Carlo Erba Strumentazione, Milan, Italy). Urinary nitrogen was converted to protein equivalents by multiplication by a factor of 6.25 (assuming that proteins contain 16% nitrogen on average) after dermal and fecal losses of nitrogen were adjusted for by the addition of 2 g nitrogen (36). Individual nitrogen recovery was ascertained by comparing urinary protein equivalents with protein intake reported in the dietary records. Thus, nitrogen recoveries >100% indicated that subjects either underreported protein intake or were in negative nitrogen balance.
 | (1) |
Lithium chloride, a salt completely excreted in urine and previously used for compliance measurements (37, 38), was added to the wheat bread during week 5 in such a concentration that subjects in the 10 MJ/d energy requirement group received 0.24 mmol lithium chloride/d. Samples of urine collections in week 5 were analyzed for lithium by flame atomic absorption spectrometry at Nova Medical Medi-Lab (Copenhagen) for evaluation of the subjects’ compliance with the test foods diet.
Fasting blood samples were collected in weeks 0 and 10 for analysis of glucose, insulin, nonesterified fatty acids (NEFAs), triacylglycerol, and total, LDL, and HDL cholesterol. All blood samples were taken from an antecubital arm vein. Blood for glucose analysis was collected in tubes containing iced EDTA prepared with sodium flouride. Blood for NEFA analysis was collected in tubes containing iced EDTA. The blood for other analyses was centrifuged for 15 min at 2800 x g and 4 °C, and serum and plasma were stored at –20 °C until they were analyzed. NEFAs, glucose, triacylglycerol, and total and HDL cholesterol were measured with the use of an enzymatic colorimetric method on a Cobas Mira Plus spectrophotometer (Roche Diagnostic Systems, F Hoffmann-La Roche, Basel, Switzerland).
Plasma NEFA concentrations were measured by using a Wako 994–75409 NEFA-C test kit (ACS-ACOD method; Wako Chemicals GmbH, Neuss, Germany) with intraassay and interassay CVs <4.5% and <4.2%, respectively. Plasma glucose concentrations were measured by using a gluco-quant Glucose/HK kit (GLU Roche/Hitachi 1447513,;Roche Diagnostics GmbH, Mannheim, Germany; 39) with intraassay and interassay CVs <2% and <4%, respectively. Serum triacylglycerol concentrations were measured by using the Triacylglycerols GPO-PAP kit (TAG Roche/Hitachi 2016648; Roche Diagnostics GmbH; 40) with intraassay and interassay CVs <2% and <3%, respectively. Serum total cholesterol concentrations were measured by using the CHOL Cholesterol CHOD-PAP kit (Roche/Hitachi 2016630; Roche Diagnostics GmbH) with intraassay and interassays CV <1.5% and <3%, respectively. Serum HDL cholesterol concentrations were measured by using the HDL Cholesterol plus kit (HDL-Cholesterol plus 1930672; Roche Diagnostics GmbH; 41) with intraassay and interassays CV <1.2% and <5.1%, respectively. Serum LDL-cholesterol concentrations were calculated by using the equation of Friedewald et al (42):
 | (2) |
Serum insulin concentrations were determined by using an enzyme-linked immunosorbent assay technique (AutoDELFIA Insulin kit B080-101; Wallac Oy, Turku, Finland) on a assay system (AutoDELFIA 1235–514; Wallac Oy; 44) with intraassay and interassay CVs <3.3% and <7.6%, respectively.
Estimates of relative insulin resistance and pancreatic ß-cell function, introduced as a homeostasis model assessment (HOMA-R and HOMA-ß, respectively) by Matthews et al (45), were calculated according to the following formulas:
 | (3) |
 | (4) |
Questionnaires
Subjects were instructed to fill in a diary every day during the 10-wk intervention. This diary gave information on feces patterns, including indications of consistency and frequency and whether the patterns were normal for the subject. Subjects were also instructed to fill out every evening visual analogue scales for sensations of general hunger, fullness, and well-being during the day and palatability of the foods eaten. Finally, subjects were instructed to give information on any medication taken or any unusual events, such as illness, or both. Dates of menstrual periods were also reported. Before the study, subjects completed the physical activity questionnaire of Baecke et al (17) and the Three-Factor Eating Questionnaire of Stunkard and Messick (46; Table 1).
Statistical analyses
Differences between groups in baseline values were analyzed with the use of Student’s unpaired t test. An individual mean of the 70 d was calculated for data on amounts of test foods eaten (self-reported compliance) and diary data on fecal patterns, appetite, palatability of foods, and general well-being. Differences between groups in these data and in lithium and nitrogen recoveries were analyzed with the use of Student’s unpaired t test. Differences between groups in week 0 to week 10 changes in fasting blood sample values, anthropometric measurements, heart rate, and blood pressure data were analyzed using analysis of covariance with baseline values as cofactors (PROC GLM procedure in SAS software, version 6.12; SAS Institute, Cary, NC). For blood lipid analysis, a further analysis of covariance with body-weight change as the second cofactor was performed. The mean changes from week 0 to week 10 within the 2 groups in fasting blood sample values, anthropometric measurements, and heart rate and blood pressure data were analyzed with the use of Student’s paired t test.
Residual plots from the weight change analysis showed a clear systematic pattern of lower residuals for high and low predicted values. Therefore, only results from the analysis of absolute body weight were included in the statistical analysis. Total body weight over time was analyzed by using repeated measurements to test the effect of diet, time, and the diet x time interaction with baseline value as cofactor (PROC MIXED procedure in SAS). Because diet x time interactions were not significant, the model was reduced. To test for diet differences in slope over time, time was removed as a class variable.
Data on total energy intake, energy density, macronutrient composition, and sucrose, starch, and dietary fiber intakes from the dietary records at baseline, week 5, and week 10 were analyzed by using a split-plot model (PROC MIXED procedure in SAS) testing diet, time, and diet x time interaction with baseline values as cofactors. When diet x time interactions were not significant, the model was reduced. In addition, the mean energy intakes for weeks 5 and 10 were calculated, and differences between groups were tested by using analysis of covariance with baseline values as cofactors. EI/BMR for the 3 dietary record periods was analyzed for differences between groups with the use of Student’s unpaired t test.
All statistical analyses were performed by using SAS software (version 6.12; SAS Institute). Results were considered significant when P < 0.05. Results are reported as means ± SEMs.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Urinary lithium recovery from wheat bread did not differ significantly between groups (LGI: 74 ± 8%; HGI: 82 ± 6%), but large variances among subjects were observed. Four samples with <85% PABA recovery were excluded from the nitrogen and lithium analyses.
Dietary record data
There was a significant decrease in energy intake with time, but there were no significant differences between groups (Table 4). Testing the combined mean energy intakes from weeks 5 and 10, with the baseline value as cofactor, did not change relation (LGI diet group: 9.0 ± 0.2 MJ/d; HGI diet group: 9.6 ± 0.3 MJ/d; P = 0.11). Mean EI/BMR for all dietary record data combined (total of 135 seven-day diaries) was 1.44 ± 0.02, which, when compared with Goldberg’s cutoff limits for 7-d dietary records (1.50 for n = 100 and 1.51 for n = 200; 95% confidence limits; 18), indicated that the energy intakes of the subjects were underreported. As for energy intake, EI/BMR decreased over time, but there was no significant difference between groups (Table 4).
|
): ie, carbohydrate, protein, and dietary fiber intakes increased and fat intakes decreased. Energy density did not differ significantly from that of the habitual diet (Table 4).
Nitrogen recovery
There were no significant differences between groups in nitrogen recovery at baseline (LGI diet group: 106 ± 4%; HGI diet group: 106 ± 4%; P = 0.94) or in week 5 (LGI diet group: 98 ± 3%; HGI diet group: 92 ± 4%; P = 0.28). However, in week 10, the nitrogen recovery was significantly higher in the LGI diet group than in the HGI diet group (LGI diet group: 109 ± 5%; HGI diet group: 96 ± 4%; P = 0.04; adjusted for changes in body weight, P = 0.05).
Diary data
Diaries, filled in on a daily basis by the subjects themselves, showed no differences between groups in fecal patterns or in ratings of fullness, hunger, well-being, or the palatability of test foods. Some subjects reported flatulence and general gastrointestinal discomfort, especially at the beginning of the study (8 subjects in the HGI diet group and 7 subjects in the LGI diet group).
Body weight and composition
There was no significant difference in body-weight changes between groups (LGI: –1.9 ± 0.5 kg; HGI: –1.3 ± 0.3 kg; P = 0.31), but body weight decreased significantly in both groups over time (Figures 1 and 2). Slopes of body-weight curves did not differ significantly between groups (P = 0.52).
|
|
). Waist-to-hip ratio and sagittal height did not change from week 0 to week 10, and no significant differences were observed between groups (Table 5).
|
).
Fasting blood samples
Plasma glucose, serum insulin concentrations, and homeostasis model assessment for relative insulin resistance and for ß cell function
No significant difference between groups was observed in fasting plasma glucose concentrations at baseline, but a significant difference was observed between groups in changes from week 0 to week 10 (Table 5). This finding was primarily due to a slight increase in the glucose concentrations in the LGI diet group.
Two fasting serum samples from week 10 (one sample from each group) were hemolyzed, and these data were excluded from the serum insulin, HOMA-R, and HOMA-ß analysis. There were no differences between groups in baseline values or in changes from week 0 to week 10 (Table 5).
Blood lipids
No significant difference between groups was observed in serum LDL-cholesterol concentrations at baseline (LGI diet group: 2.49 ± 0.18 mmol/L; HGI diet group: 2.63 ± 0.18 mmol/L), but a significant (P = 0.02) difference between groups was observed in the changes from week 0 to week 10: a decrease in the LGI diet group and a small increase in the HGI diet group (Figure 3). This difference was still significant after adjustment for body-weight changes (P = 0.03). No significant difference was observed in serum total cholesterol concentrations at baseline (LGI diet group: 4.60 ± 0.21 mmol/L; HGI diet group: 4.79 ± 0.22 mmol/L). However, a tendency (P = 0.06) to a larger decrease from week 0 to week 10 was seen in the LGI diet group than in the HGI diet group (Figure 3). This difference disappeared after adjustment for body-weight changes (P = 0.11). No significant (P = 0.50) differences between groups were observed in serum HDL-cholesterol concentration at baseline (LGI diet group: 1.54 ± 0.06 mmol/L; HGI diet group: 1.63 ± 0.07 mmol/L) or in changes from week 0 to week 10 (Figure 3). The ratio of LDL to HDL cholesterol did not differ significantly between groups at baseline (LGI diet group: 1.66 ± 0.13; HGI diet group: 1.67 ± 0.13), but from week 0 to week 10, a tendency (P = 0.0548) was seen for a small decrease in the LGI diet group and an increase in the HGI diet group. This tendency was still seen after adjustment for body-weight changes (P = 0.06). The ratio of total to HDL cholesterol showed no significant difference between groups at baseline (LGI diet group: 3.05 ± 0.16; HGI diet group: 3.01 ± 0.15 mmol/L) or in changes from week 0 to week 10 (P = 0.11).
|
). Moreover, no significant differences between groups were observed in plasma NEFA concentrations at baseline (LGI diet group: 457 ± 31 µmol/L; HGI diet group: 479 ± 38 µmol/L) or in changes from week 0 to week 10 (Figure 3). |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The members of the LGI diet group did not lower their energy intakes significantly more than did those of the HGI diet group. Therefore, our data do not support the contention that LGI diets are more satiating and result in lower ad libitum energy intake than do HGI diets (5, 7, 47). In both the LGI and HGI diet groups, ad libitum energy intake decreased over time. This is most likely due to the fact that the test foods were richer in carbohydrates and dietary fiber and therefore were more satiating than were the subjects’ habitual foods (48, 49). Furthermore, the fact that subjects were included in a trial could have the effect of decreasing energy intake. The decrease in energy intake observed in these subjects resulted in a decrease in body weight and fat mass in both groups. Even though subjects in the LGI diet group lost a total of 1.9 kg compared with 1.3 kg lost by the HGI diet group, this difference was not significant. In comparison, a 5-wk crossover study with 11 overweight men found body weight to be 0.8 kg lower and energy intake 1 MJ lower after an LGI than after an HGI diet, but, as in the present study, none of those differences were significant (8). The results from the 5-wk study could, however, be due to higher fiber content in the LGI diet than in the HGI diet. In contrast, a study of overweight children found significantly greater weight loss after 4 mo intervention in a pediatric obesity program comparing LGI diets with the traditional recommendations, but many other factors beside the GI varied between groups in that study (50). On the other hand, a 2-wk crossover study with 18 normal-weight women showed a lower ad libitum energy intake with an HGI (high-starch) diet than with an LGI (high-sucrose) diet and a weight loss (0.7 kg) with the HGI diet but a weight gain (0.1 kg) with the LGI diet (51). However, in that study, the high-starch diet had a higher fiber content than did the high-sucrose diet, which could have influenced the results. In addition, a 4-mo study of 24 subjects with impaired glucose tolerance found a significantly smaller weight loss after an LGI diet (0.19 kg) than after an HGI diet (0.49 kg) (10). Other longer-term studies were designed as weight-maintenance or energy-restriction studies, and neither design is appropriate for ascertaining effects on appetite and voluntary energy intake in the long term (6).
With reference to the weight-change curves in the present study, it could be speculated that a longer study period would have found some significant differences. The modest weight losses observed, however, suggest that more study power in the form of an increased number of subjects would probably be needed to detect significant differences between groups.
In the present study, dual-energy X-ray absorptiometric scanning found no differences between groups in changes in fat mass or fat-free mass. In contrast, the abovementioned ad libitum study with 11 overweight men showed a lower fat mass and a tendency to a higher fat-free mass after 5 wk of an LGI diet than after 5 wk of an HGI diet (8). That HGI diets should have an adverse effect on body composition has been suggested by Agus et al (52) to be due to the fact that proteolytic counterregulatory hormones lead to catabolism of lean body mass. This theory is supported by their findings of a more negative nitrogen balance with consumption of an HGI diet than with consumption of an LGI diet (52). This finding was, however, not supported by our study, in which data instead suggested a more negative nitrogen balance in the LGI diet group than in the HGI diet group.
No differences between the LGI and HGI diet groups were seen in risk markers for type 2 diabetes, although fasting glucose increased after the LGI diet. This small increase was unexpected, and, when the individual data were evaluated, no odd values were found. The finding may therefore be a result of many comparisons (type I error). It could perhaps also be explained by a slight hemoconcentration in the LGI diet group caused by water retention due to more slowly digestible starch and more slowly available glucose. This hypothesis is supported by the observation that the reduction in fat-free mass in the LGI diet group was less than would be expected on the basis of the reduction in body weight. The lack of differences observed between groups in fasting insulin concentrations, relative insulin sensitivity (HOMA-R), or ß-cell function (HOMA-ß) suggests that, for these subjects, the type of carbohydrate is not important in relation to insulin sensitivity. However, evaluation of possible effects on insulin sensitivity would have required the use of clamp procedure, which was not performed in this study.
A significant difference between groups was seen in LDL cholesterol. Thus, 10 wk of the LGI diet resulted in an 10% decrease in LDL cholesterol, whereas a small increase (2%) was seen after 10 wk of the HGI diet. This difference was still observed when LDL cholesterol was expressed relative to HDL cholesterol. In addition, a tendency to a difference between groups in total cholesterol was seen: 7.2% and 1.9% decreases in the LGI and HGI diet groups, respectively. Our results, therefore, support data from previous studies, most of which were performed with diabetic or hyperlipidemic subjects, showing lower LDL cholesterol after LGI diets than after HGI diets (53). The only previous study with healthy subjects showed a 15% decrease in total cholesterol after 2 wk of an energy-fixed LGI diet but only a 2% decrease after a similar HGI diet (54). No increases or differences were seen between our 2 groups in triacylglycerol or HDL-cholesterol values. Our results do, therefore, not support the contention that a carbohydrate-rich diet increases plasma triacylglycerol concentrations (3, 55).
The test foods given to subjects during the present study were well matched in macronutrient composition, total energy, energy density, and dietary fiber. The difference in GI, calculated from in vitro measurements of HI, was 24.3. In comparison, Jenkins et al (4) found a mean difference in GI of 19.6 (range: 1.2–41) when they reviewed 19 GI studies. A major difference between the test foods designed for our study was the difference in food structure. Järvi et al (56) designed 2 GI diets by manipulating the structure of starchy foods without altering other characteristics of the foods. Their study had an average GI difference between diets of 26.2. The HI method used in the present work and in the study by Järvi et al was evaluated specifically for different pasta, rice, and bread types and was shown to predict GI with good accuracy (20).
In conclusion, this study does not support the contention that low-fat LGI diets are more beneficial than low-fat HGI diets with respect to appetite and body weight regulation when evaluated in an ad libitum context over 10 wk. However, it does support previous findings of a beneficial effect of LGI diets on risk factors for IHD. Further long-term studies, preferentially 6–12 mo, are needed to substantiate these findings.
|
ACKNOWLEDGMENTS |
|---|
AR was responsible for the first study protocol. BS, IK-M, and AR were responsible for conducting the trial and for data collection. BS was responsible for data analysis and drafted the first manuscript with supervision from AR and AF. AA was medical counselor for the project. IB and HE were responsible for the hydrolysis index analysis of test foods. All authors contributed to the development of the final study design and interpretation of the results. SV and VL are employed at Danone Vitapole, France. None of the other authors had any conflict of interest.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. A Parnell and R. A Reimer Weight loss during oligofructose supplementation is associated with decreased ghrelin and increased peptide YY in overweight and obese adults Am. J. Clinical Nutrition, June 1, 2009; 89(6): 1751 - 1759. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Metwally, S. Amer, T.C. Li, and W.L. Ledger An RCT of metformin versus orlistat for the management of obese anovulatory women Hum. Reprod., April 1, 2009; 24(4): 966 - 975. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. B. Scribner, D. B. Pawlak, C. M. Aubin, J. A. Majzoub, and D. S. Ludwig Long-term effects of dietary glycemic index on adiposity, energy metabolism, and physical activity in mice Am J Physiol Endocrinol Metab, November 1, 2008; 295(5): E1126 - E1131. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Boucher, G. A. Benson, S. Kovarik, B. Solem, and J. J. VanWormer Current Trends in Weight Management: What Advice Do We Give to Patients? Clin. Diabetes, July 1, 2008; 26(3): 115 - 120. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Marsh and J. Brand-Miller State of the Art Reviews: Glycemic Index, Obesity, and Chronic Disease American Journal of Lifestyle Medicine, April 1, 2008; 2(2): 142 - 150. [Abstract] [PDF]
|
|
|
|
|
|
J.-P. Chaput, A. Tremblay, E. B Rimm, C. Bouchard, and D. S Ludwig A novel interaction between dietary composition and insulin secretion: effects on weight gain in the Quebec Family Study Am. J. Clinical Nutrition, February 1, 2008; 87(2): 303 - 309. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Jensen, B. Sloth, I. Krog-Mikkelsen, A. Flint, A. Raben, T. Tholstrup, N. Brunner, and A. Astrup A low-glycemic-index diet reduces plasma plasminogen activator inhibitor-1 activity, but not tissue inhibitor of proteinases-1 or plasminogen activator inhibitor-1 protein, in overweight women Am. J. Clinical Nutrition, January 1, 2008; 87(1): 97 - 105. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Livesey, R. Taylor, T. Hulshof, and J. Howlett Glycemic response and health a systematic review and meta-analysis: the database, study characteristics, and macronutrient intakes Am. J. Clinical Nutrition, January 1, 2008; 87(1): 223S - 236S. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Weight Loss Requires Drop in Calories, Not Low GI DOC News, November 1, 2007; 4(11): 4 - 4. [Full Text]
|
|
|
|
|
|
J. Brand-Miller Effects of glycemic load on weight loss in overweight adults Am. J. Clinical Nutrition, October 1, 2007; 86(4): 1249 - 1250. [Full Text] [PDF]
|
|
|
|
|
|
R. Sichieri, A. S Moura, V. Genelhu, F. Hu, and W. C Willett An 18-mo randomized trial of a low-glycemic-index diet and weight change in Brazilian women Am. J. Clinical Nutrition, September 1, 2007; 86(3): 707 - 713. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Boucher, G. A. Benson, S. Kovarik, B. Solem, and J. J. VanWormer Current Trends in Weight Management: What Advice Do We Give to Patients? Diabetes Spectr, July 1, 2007; 20(3): 153 - 158. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. B Levitan, M. A Mittleman, N. Hakansson, and A. Wolk Dietary glycemic index, dietary glycemic load, and cardiovascular disease in middle-aged and older Swedish men Am. J. Clinical Nutrition, June 1, 2007; 85(6): 1521 - 1526. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. B. Ebbeling, M. M. Leidig, H. A. Feldman, M. M. Lovesky, and D. S. Ludwig Effects of a Low-Glycemic Load vs Low-Fat Diet in Obese Young Adults: A Randomized Trial JAMA, May 16, 2007; 297(19): 2092 - 2102. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. C Maki, T. M Rains, V. N Kaden, K. R Raneri, and M. H Davidson Effects of a reduced-glycemic-load diet on body weight, body composition, and cardiovascular disease risk markers in overweight and obese adults Am. J. Clinical Nutrition, March 1, 2007; 85(3): 724 - 734. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A Pereira Reply to EJ Mayer-Davis Am. J. Clinical Nutrition, March 1, 2007; 85(3): 922 - 923. [Full Text] [PDF]
|
|
|
|
 |
|
 J. Brand-Miller, J. McMillan-Price, and P. Petocz Glycemic Load and Cardiovascular Risk--Reply Arch Intern Med, January 22, 2007; 167(2): 206 - 207. [Full Text] [PDF]
|
|
|
|
|
|
D. D Motton, N. L Keim, F. A Tenorio, W. F Horn, and J. C Rutledge Postprandial monocyte activation in response to meals with high and low glycemic loads in overweight women Am. J. Clinical Nutrition, January 1, 2007; 85(1): 60 - 65. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A Pereira Weighing in on glycemic index and body weight. Am. J. Clinical Nutrition, October 1, 2006; 84(4): 677 - 679. [Full Text] [PDF]
|
|
|
|
|
|
H. Hare-Bruun, A. Flint, and B. L Heitmann Glycemic index and glycemic load in relation to changes in body weight, body fat distribution, and body composition in adult Danes. Am. J. Clinical Nutrition, October 1, 2006; 84(4): 871 - 879. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. J Stevenson, C. Williams, L. E Mash, B. Phillips, and M. L Nute Influence of high-carbohydrate mixed meals with different glycemic indexes on substrate utilization during subsequent exercise in women. Am. J. Clinical Nutrition, August 1, 2006; 84(2): 354 - 360. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. McMillan-Price, P. Petocz, F. Atkinson, K. O'Neill, S. Samman, K. Steinbeck, I. Caterson, and J. Brand-Miller Comparison of 4 diets of varying glycemic load on weight loss and cardiovascular risk reduction in overweight and obese young adults: a randomized controlled trial. Arch Intern Med, July 24, 2006; 166(14): 1466 - 1475. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Valtuena, N. Pellegrini, D. Ardigo, D. Del Rio, F. Numeroso, F. Scazzina, L. Monti, I. Zavaroni, and F. Brighenti Dietary glycemic index and liver steatosis Am. J. Clinical Nutrition, July 1, 2006; 84(1): 136 - 142. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L. Neuhouser, L. F. Tinker, C. Thomson, B. Caan, L. V. Horn, L. Snetselaar, L. M. Parker, R. E. Patterson, R. Robinson-O'Brien, S. A. A. Beresford, et al. Development of a Glycemic Index Database for Food Frequency Questionnaires Used in Epidemiologic Studies J. Nutr., June 1, 2006; 136(6): 1604 - 1609. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Murakami, S. Sasaki, Y. Takahashi, H. Okubo, Y. Hosoi, H. Horiguchi, E. Oguma, and F. Kayama Dietary glycemic index and load in relation to metabolic risk factors in Japanese female farmers with traditional dietary habits Am. J. Clinical Nutrition, May 1, 2006; 83(5): 1161 - 1169. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Ma, Y. Li, D. E. Chiriboga, B. C. Olendzki, J. R. Hebert, W. Li, K. Leung, A. R. Hafner, and I. S. Ockene Association between Carbohydrate Intake and Serum Lipids. J. Am. Coll. Nutr., April 1, 2006; 25(2): 155 - 163. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. C.G. Alfenas and R. D. Mattes Influence of Glycemic Index/Load on Glycemic Response, Appetite, and Food Intake in Healthy Humans Diabetes Care, September 1, 2005; 28(9): 2123 - 2129. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Brand-Miller Optimizing the cardiovascular outcomes of weight loss Am. J. Clinical Nutrition, May 1, 2005; 81(5): 949 - 950. [Full Text] [PDF]
|
|
|
|
|
|
C. B Ebbeling, M. M Leidig, K. B Sinclair, L. G Seger-Shippee, H. A Feldman, and D. S Ludwig Effects of an ad libitum low-glycemic load diet on cardiovascular disease risk factors in obese young adults Am. J. Clinical Nutrition, May 1, 2005; 81(5): 976 - 982. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M Strik and C J. Henry The role of a low-glycemic-index diet in the management of obesity Am. J. Clinical Nutrition, April 1, 2005; 81(4): 940 - 941. [Full Text] [PDF]
|
|
|
|
|
|
B. Sloth, I. Krog-Mikkelsen, A. Flint, I. Tetens, A. Astrup, A. Raben, and H. Elmstahl Reply to CM Strik and CJ Henry Am. J. Clinical Nutrition, April 1, 2005; 81(4): 941 - 941. [Full Text] [PDF]
|
|
|
|
|
|
J. Brand-Miller Glycemic index and body weight Am. J. Clinical Nutrition, March 1, 2005; 81(3): 722 - 723. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
