Acute effects on insulin sensitivity and diurnal metabolic profiles of a high-sucrose compared with a high-starch diet1-3
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Acute effects on insulin sensitivity and diurnal metabolic profiles of
a high-sucrose compared with a high-starch diet1–3
Mark E Daly, Catherine Vale, Mark Walker, Alison Littlefield, K George MM Alberti, and John C Mathers
ABSTRACT Decreased insulin sensitivity is associated with Obesity and exercise are associated with decreased and
diabetes mellitus, ischemic heart disease, and hypertension, both increased insulin sensitivity, respectively (3, 4). Himsworth (5)
independently and in association as what is called the metabolic was one of the first to explore the effects of diet on insulin sen-
syndrome. Although the negative effects of obesity, sedentary sitivity, contrasting high-carbohydrate and low-carbohydrate
lifestyles, and high-fat diets on insulin sensitivity are well estab- diets. Since then animal work has clearly shown the capacity of
lished, the influence of type and quantity of dietary carbohydrate high-fat, high-sucrose, or high-fructose diets to reduce insulin
Downloaded from ajcn.nutrition.org by guest on February 4, 2015
is more controversial. This study aimed to assess the acute (24 h) sensitivity (6–13), but this has not been evident at the lower
effects of a high-sucrose compared with a high-starch diet on dietary concentrations of sucrose or fructose usually given in
insulin sensitivity and to identify changes in blood metabolites human studies. Some human studies have shown a positive effect
that might lead to altered insulin sensitivity. Eight healthy adults (improvement in insulin sensitivity) of a high-complex-carbohy-
consumed high-sucrose or high-starch diets (50% of dietary drate, low-fat diet (14, 15), but others have shown no change
energy) in a randomized, crossover trial. Insulin sensitivity was with such a diet (16, 17). Work on the type of dietary carbohy-
assessed by a short insulin tolerance test the following morning. drate has also been inconclusive, with several studies showing no
No differences were detected in insulin sensitivity, either for glu- effect on insulin sensitivity (18–22), some showing an adverse
cose metabolism [Kittglucose (the rate constant for the decline in effect of sucrose or fructose reflected by a rise in fasting serum
blood glucose concentrations) for sucrose diet = 3.86 %/min, for insulin concentrations (23, 24), and one showing an adverse
starch diet = 3.72%/min; pooled SEM = 0.23] or for lipid metab- effect of fructose as measured by the insulin tolerance test (25).
olism [KittNEFA (the rate constant for the decline in blood fatty By contrast, Koivisto and Yki-Jarvinen (26) reported a beneficial
acid concentrations) for sucrose diet = 12.9%/min, for starch diet effect of fructose on insulin sensitivity and a recent study
= 11.4%/min; pooled SEM = 1.18]. Profiles for blood glucose showed an increase in insulin sensitivity with a diet with a high
and serum insulin concentrations revealed higher peaks and glycemic index (27), but only with the high-dose insulin step of
lower troughs with the high-sucrose diet whereas area under the a hyperinsulinemic, euglycemic clamp.
curve for glucose was higher with the high-starch diet (6780 ± The present study was designed to establish whether insulin
245 mmol · L/min) than with the high-sucrose diet (6290 ± 283 sensitivity is affected acutely by major manipulation of the type
mmol · L/min) (P< 0.001). Plasma fatty acid concentrations of dietary carbohydrate, and whether the metabolic profiles asso-
showed a late postprandial rise with the sucrose-rich diet relative ciated with this intervention could provide information on possi-
to the starch-rich diet, which was mirrored with a fractionally ble mechanisms underlying altered sensitivity. Previous work
later peak in triacylglycerol concentrations. Am J Clin Nutr has been done on the postprandial metabolic profiles of sucrose
1998;67:1186–96. and fructose, both in healthy individuals and patients with dia-
betes. Coulston et al (28) examined the effects of a sucrose-rich
KEY WORDS Insulin sensitivity, dietary carbohydrate, diet on glucose, insulin, and triacylglycerol profiles in people
sucrose, starch, glucose metabolism, fatty acids, triacylglycerol,
lipid metabolism
1
From the Human Nutrition Research Centre, the Department of Biolog-
INTRODUCTION ical and Nutritional Sciences, the Human Diabetes and Metabolism Research
Decreased insulin sensitivity is recognized as a major meta- Centre, and the Department of Medicine, University of Newcastle upon Tyne,
bolic feature of type 2 diabetes and is one of the earliest detectable United Kingdom.
2
Supported by the Ministry of Agriculture, Fisheries and Food (project
abnormalities in some people who go on to develop type 2 dia-
no. ANO309); the British Diabetic Association; and Novo Nordisk Industries.
betes. An association among cardiovascular risk factors (type 2 3
Address reprint requests to ME Daly, Human Nutrition Research Centre,
diabetes, hyperlipidemia, glucose intolerance, and obesity) was University of Newcastle upon Tyne, Royal Victoria Infirmary, Newcastle
described by Avogaro and Crepaldi (1); Reaven (2) extended this upon Tyne, NE1 4LP United Kingdom. E-mail: m.e.daly@ncl.ac.uk.
with a description of the metabolic syndrome, or syndrome X, in Received June 17, 1997.
which decreased insulin sensitivity is a key feature. Accepted for publication November 17, 1997.
1186 Am J Clin Nutr 1998;67:1186–96. Printed in USA. © 1998 American Society for Clinical NutritionINSULIN SENSITIVITY AND DIURNAL METABOLIC PROFILES 1187
with diabetes and found small rises in the postprandial triacyl- jects received the high-starch diet first and four received the
glycerol areas under the curve with this diet compared with a high-sucrose diet first. Subjects were admitted at 0700, having
high-starch diet (in which other macronutrient contributions fasted from 2200 the previous evening. Alcohol and strenuous
were fixed). Work has also been done comparing fructose-rich exercise were avoided for 24 h before each experimental period.
with sucrose-rich diets in both patients with diabetes (19) and Study periods were separated by a minimum of 1 wk (men), with
those without (29), but no information on postprandial triacyl- 1 mo for females to keep as close as possible to a fixed point in
glycerol profiles was reported. Raised triacylglycerol concentra- the menstrual cycle because concentrations of blood lipids can
tions may have some association with altered insulin sensitivity; vary with stage of the menstrual cycle (31, 32). Each dietary
therefore, the present study sought to extend the knowledge in period was followed immediately by an insulin tolerance test.
this area by examining in detail a 24-h metabolic profile of glu- Frequent measurements of blood glucose, serum insulin, plasma
cose, insulin, fatty acids (NEFAs), triacylglycerols, and carbohy- NEFA, serum triacylglycerol, and other blood metabolite (pyru-
drate metabolite concentrations in healthy subjects consuming a vate, lactate, and glycerol) concentrations were made during
sucrose-rich compared with a starch-rich diet. each 24-h study.
Percentage body fat was calculated by the equations of Siri
(33) from estimates of body density derived from skinfold thick-
SUBJECTS AND METHODS ness measurements (using a Holtain/Tanner-Whitehouse skin
Subjects caliper; Holtain, Crosswell, United Kingdom) at four separate
sites (34). Body composition was also estimated by bioelectrical
The experimental protocol was approved by the joint ethics impedance analysis (Holtain Body Composition Analyzer).
committees of Newcastle and North Tyneside Health authorities,
Insulin sensitivity and metabolic profiles
University of Newcastle upon Tyne, and University of Northum-
Downloaded from ajcn.nutrition.org by guest on February 4, 2015
bria at Newcastle, and each subject gave his or her informed, Insulin sensitivity was assessed by using a modified insulin
written consent to participate. All studies were conducted in the tolerance test (35). After the injection of 0.05 units of insulin per
Wellcome Research Laboratories, Royal Victoria Infirmary, kg body mass, blood samples were taken for measurement of
Newcastle upon Tyne. plasma glucose and NEFA concentrations at 1-min intervals
Eight healthy, weight-stable volunteers (four men and four from 3 to 15 min. Arterial blood samples were taken from a ret-
women) were recruited from the student and staff populations of rograde intravenous cannula in a dorsal hand vein, 15 min before
the University of Newcastle upon Tyne and from the staff of the the test. This hand was kept in a heated box at 55 °C to facilitate
Royal Victoria Infirmary, Newcastle upon Tyne. None had dia- arterial blood sampling. Rate constants for the decrease in sub-
betes mellitus (or a first-degree relative with diabetes), ischemic strate concentrations (Kitt) were derived from plasma glucose
heart disease, hypertension, or any other disease associated with concentrations at 3–15 min and for plasma NEFA concentrations
altered insulin sensitivity. None were taking any drugs known to at 6–15 min, by regression analysis of the natural logarithm of
alter insulin sensitivity or to affect carbohydrate or lipid metab- glucose and NEFA concentrations, respectively. A typical insulin
olism. All were nonsmokers and had habitual alcohol intakes of tolerance test analysis, with the log-transformed curves for glu-
< 21 units per week [one unit = 8 g alcohol (30)]. cose and NEFA concentrations, is shown in Figure 1. The R2 val-
ues for the regression lines from which the rate constants were
Experimental protocol
derived are also shown.
Each subject (n = 8) took part in two experimental periods of Blood samples for the metabolic profile were taken over a 22-h
24 h each in a randomized crossover design, such that four sub- period from an antecubital vein at 30-min intervals for 2 h after
FIGURE 1. An example of the modified insulin tolerance test. After the injection of 0.05 units of insulin per kg body mass, plasma glucose and
fatty acid (NEFA) concentrations were measured at 1-min intervals from 3 to 15 min. Shown are the log-transformed curves for glucose and NEFA,
with R2 values for the regression lines from which the rate constants were derived.1188 DALY ET AL
each of the four meals and then hourly, except overnight when assessed with 99% CIs about the zero difference, with a higher
sampling was every 2 h. Blood glucose (and plasma glucose for CI used to examine the statistical effects of multiple testing. Glu-
the insulin tolerance test) was analyzed with a glucose analyzer cose area under the curve analyses were calculated by multiply-
(Yellow Springs International, Yellow Springs, OH) with use of ing the mean glucose concentration at a pair of adjacent time
the glucose oxidase method (interassay CV: 1.7%). Serum insulin points by the intervening time interval, and summing each of
was estimated by using an enzyme-linked immunosorbent assay these products to give the total area under curve. Data are pre-
specific for insulin (intraassay CV: 2.9%; Dako Diagnostics Ltd, sented as means ± SEMs or SDs as appropriate.
Ely, United Kingdom), plasma NEFAs were estimated by using an
enzymatic colorimetric method (intraassay CV: 3.2%; Wako
Chemicals GmbH, Neuss, Germany), and triacylglycerols were RESULTS
estimated by enzymatic hydrolysis of triacylglycerols with subse- Subject characterization
quent enzymatic assay for glycerol (intraassay CV: 3.9%; Wako
Chemicals GmbH). Blood metabolites [pyruvate (intraassay CV: The volunteers for this study were healthy, young (mean age:
4.9%), lactate (intraassay CV: 3.5%), and glycerol (intraassay CV: 25 y; range: 20–31 y) men and women with body mass indexes
4.3%)] were measured by using a centrifugal analyzer fitted with (in kg/m2) < 25 (Table 3). Body fat content as estimated by bio-
a fluorimetric attachment (36). electrical impedance analysis correlated strongly with skinfold
thickness measurements (r = 0.96), but values for percentage
Diets
body fat wereINSULIN SENSITIVITY AND DIURNAL METABOLIC PROFILES 1189
TABLE 2
Macronutrient contribution of each meal provided by the high-starch and high-sucrose diets1
Carbohydrate2 Carbohydrate2 Fat Fat Protein Protein
% of energy g/d % of energy g/d % of energy g/d
Breakfast
High-starch meal 10 48.0 6 12.9 2.4 10.9
High-sucrose meal 11 52.8 5.5 11.8 1.5 6.8
Lunch
High-starch meal 17.8 85.5 12 25.8 3.8 17.2
High-sucrose meal 19 91.3 12.3 26.4 4 18.1
Dinner
High-starch meal 18.9 90.8 11.9 25.6 3.7 16.8
High-sucrose meal 18.8 90.3 11.8 25.3 3 13.6
Supper
High-starch meal 7.5 36.0 4.7 10.1 1 4.5
High-sucrose meal 7.6 36.5 3.9 8.4 1.6 7.3
Total
Starch diet — 260 — 74 — 49
Sucrose diet — 271 — 72 — 46
1
Both diets provided a total of 8.4 MJ/d.
2
Available carbohydrate (39).
Downloaded from ajcn.nutrition.org by guest on February 4, 2015
than amounts in the average UK diet and met current guidelines Figure 3). With the high-sucrose diet, blood glucose rose more
produced by the UK Department of Health (40). The subjects’ rapidly to a higher peak after each of the four meals. Serum insulin
habitual diets were high in carbohydrate, low in total fat, low in concentrations also rose more rapidly after each meal with the
saturated fat, and moderate in protein intake (Table 4). We were high-sucrose diet. After these postprandial peaks, blood glucose
able to validate the food diaries by measuring urinary nitrogen concentrations fell more rapidly with the high-sucrose diet, dipping
for seven of eight volunteers, and there was a significant below the fasting value after each meal except supper. This is in
although modest correlation (r = 0.49) between mean daily nitro- contrast with results for the high-starch diet, with which mean
gen intake estimated by the food diaries and urinary nitrogen blood glucose concentrations never went below the mean fasting
excretion measured over a single 24-h period. value. The area under the glucose curve was significantly greater
for the high-starch diet than for the high-sucrose diet (x– ± SD: 6780
Insulin sensitivity ± 245 compared with 6290 ± 283 mmol · L/min; P < 0.001).
No significant differences were detected in insulin sensitivity Consumption of a high-sucrose meal substantially suppressed
as assessed by the modified insulin tolerance test for glucose plasma NEFA concentrations (Figure 4), which returned to fast-
metabolism (Kittglucose = 3.86%/min and 3.72%/min for the high- ing levels before the next meal, except after dinner, for which the
sucrose and high-starch diets, respectively; pooled SEM: 0.23). interval before supper was relatively short (3 h). The high-starch
There were also no significant differences for lipid metabolism breakfast produced a response similar to that seen with the high-
(KittNEFA = 12.9%/min and 11.4%/min, respectively; pooled sucrose breakfast, but with subsequent high-starch meals
SEM: 1.18). changes in plasma NEFA concentrations were less marked.
Triacylglycerol concentrations rose gradually with both diets
Metabolic profiles after breakfast, with a further increase in the first hour after lunch
There were marked differences in the blood glucose and serum (Figure 5). After this time point (1300), however, triacylglycerol con-
insulin concentration profiles between the diets (Figure 2 and centrations began to diverge. With the high-starch diet, triacylglycerol
TABLE 4
TABLE 3 Habitual daily dietary intakes of energy and macronutrients of eight
Anthropometric characteristics and fasting blood lipid concentrations of volunteers
the volunteers1
Percentage of Percentage of
Men (n = 4) Women (n = 4) Intake total energy nonethanol energy
Age (y) 25.5 ± 6.4 23.8 ± 1.7 g/d (MJ/d) % %
BMI (kg/m2) 23.4 ± 1.4 21.7 ± 1.3 Energy (10.1) — —
Body fat (%) Protein 78 13.3 14.1
Skinfold thickness 18.6 ± 2.3 28.1 ± 4.2 Carbohydrate 305 48.6 51.5
Impedance 21.2 ± 5.2 30.1 ± 4.1 Starch 180 28.8 30.7
Waist-to-hip ratio 0.84 ± 0.06 0.74 ± 0.05 Sucrose 63 9.8 10.4
Serum total cholesterol (mmol/L) (4.01 ± 0.45)2 Fat 86 32.1 34.4
HDL cholesterol (mmol/L) (1.32 ± 0.19) Saturated fat 28 10.3 11.1
Serum triacylglycerol (mmol/L) (0.73 ± 0.19) Nonstarch 19 — —
1–
x ± SD. polysaccharides
2
Pooled value for men and women in parentheses. Alcohol 20 6.0 —1190 DALY ET AL
Downloaded from ajcn.nutrition.org by guest on February 4, 2015
FIGURE 2. Diurnal profile of blood glucose concentrations with the high-starch and high-sucrose diets. x– ± SEM; n = 8.
concentrations gradually declined toward fasting values. With the high- tivity is affected acutely by major manipulation of the dietary
sucrose diet, concentrations continued to climb to a peak concentration carbohydrate source and whether the metabolic profiles associ-
at 1730, 0.5 h after dinner. Triacylglycerol concentrations then rapidly ated with such an intervention could provide information on pos-
declined until supper (2000), which provoked a further peak at 2200, sible mechanisms underlying the altered sensitivity.
after which concentrations gradually returned to fasting values at 0600. The acute effects of feeding high-sucrose compared with high-
Glycerol concentrations rose higher at midnight with the sucrose diet, starch diets on insulin sensitivity in healthy young adults are
but by 0600 concentrations with both diets were similar (Figure 6). clear. No significant differences between diets in insulin-stimu-
Glycolytic products showed particularly marked differences lated glucose or NEFA clearance rates were detected. The modi-
between diets. Blood lactate and pyruvate concentrations peakedINSULIN SENSITIVITY AND DIURNAL METABOLIC PROFILES 1191
FIGURE 4. Diurnal profile of fatty acid (NEFA) concentrations with the high-starch and high-sucrose diets. x– ± SEM; n = 8.
Downloaded from ajcn.nutrition.org by guest on February 4, 2015
between diets [by paired t testing of geometrically transformed although the glycemic response curve was retarded after a slowly
data (Table 5)]. This finding is consistent with the lack of effect absorbed breakfast, this was offset by a greater glucose response
of these diets on insulin sensitivity. There were minor differences curve after lunch. This likely reflects an extended fed state after
in the macronutrient compositions of the two diets (Table 2), but the slowly absorbed breakfast before the second meal. In the
these were unlikely to have been sufficient to affect the results. present study we tried to replicate a more normal eating pattern
Since Crapo et al’s (43) early work on postprandial glucose and thus examine effects in both the fed and fasted states.
and insulin responses to different carbohydrates, there have been Most of the characteristic metabolic effects of high-sucrose
many reports on postprandial responses to single test meals diets (ie, those not shared by high-starch diets), either immedi-
(44–47). However, an inherent problem of a single test meal is ately after the consumption of a test meal or with longer-term
that the interpretation of the data is restricted to effects of the exposure to sucrose-rich diets, are thought to be caused by the
food after fasting, so that possible later effects are ignored. Other fructose component of these diets (50). After absorption, a high
workers have sought to examine the possible effects when more proportion of fructose is taken up by the liver, with low concen-
than one test meal is used. Ercan et al (48) used an identical two- trations being detected in the blood after sucrose or fructose con-
meal study, with one meal consumed in the fasting state and one sumption. Hepatic fructokinase, a highly active ketohexokinase
later in the fed state. One observation was that the glucose with a strong specificity for fructose, catalyzes the phosphoryla-
response curve was greater after the second meal; thus, tion of fructose to fructose-1-phosphate. This step bypasses
responses in the fed and fasted states were different. However, phosphofructokinase, one of the key regulatory enzymes in gly-
such differences after multiple meals need not only reflect the colysis, which explains the rapid metabolism of fructose
differing extremes of the fed or fasted state. Nestler et al (49) compared with glucose by the liver. Because pyruvate kinase is
used breakfasts of differing rates of absorption and found that stimulated by fructose-1-phosphate, there is a strong drive for
FIGURE 5. Diurnal profile of serum triacylglycerol concentrations with the high-starch and high-sucrose diets. x– ± SEM; n = 8.1192 DALY ET AL
FIGURE 6. Diurnal profile of blood glycerol concentrations with the high-starch and high-sucrose diets. x– ± SEM; n = 8.
Downloaded from ajcn.nutrition.org by guest on February 4, 2015
production of pyruvate and then other metabolic intermediates hepatocytes in vitro (52). Malonyl-CoA further affects lipogene-
such as lactate, acetyl-CoA, and malonyl-CoA. Experimental sis through inhibition of carnitine O-palmitoyltransferase, the
work supports this theory of rapid metabolism of fructose. rate-limiting enzyme in the oxidation of fatty acids, shifting the
Increased production of lactate and pyruvate has been shown balance in favor of esterification. Thus, there are three mecha-
with the acute administration of fructose. Brundin and Wahren nisms by which fructose can increase triacylglycerol secretion
(51) found large differences in lactate and pyruvate concentra- from the liver; indeed, liver perfusion experiments in rats (53)
tions after the acute ingestion of a 75-g fructose load compared with fructose and insulin have shown increased VLDL secretion.
with a 75-g glucose load in human volunteers. Fructose is also involved in gluconeogenesis, but this involve-
Fructose is also known to have acute effects on lipid metabo- ment varies between the fed and fasting states, with direct glu-
lism by inducing a rapid production of the precursors necessary coneogenesis from fructose predominating over production via
for lipid synthesis. The glycolysis product dihydroxyacetone lactate in the fasting state. Experiments on catheterized humans
phosphate is readily converted to glycerol-3-phosphate, which is suggest that in the fasting state up to 66% of fructose is con-
required for the esterification of long-chain acyl-CoA in triacyl- verted into glucose (54).
glycerol synthesis. Acetyl-CoA, derived from pyruvate, is con- Changes in blood concentrations of, and especially in areas
verted to malonyl-CoA, which is then synthesized into long- under the curve for, metabolites and hormones have been widely
chain fatty acids. Fructose again differs from glucose at this interpreted as reflections of changes in the rates of appearance or
point because high lactate and pyruvate concentrations in the disappearance of these metabolites and hormones from the
liver (seen with fructose ingestion) promote the formation of bloodstream. However, changes in the rate of flux of a substance
malonyl-CoA, and the capacity of high concentrations of these are not necessarily reflected in altered circulating concentra-
glycolytic products to increase malonyl-CoA has been shown in tions, so the latter must be interpreted with care. In the context of
FIGURE 7. Diurnal profile of blood pyruvate concentrations with the high-starch and high-sucrose diets. x– ± SEM; n = 8.INSULIN SENSITIVITY AND DIURNAL METABOLIC PROFILES 1193
FIGURE 8. Diurnal profile of blood lactate concentrations with the high-starch and high-sucrose diets. x– ± SEM; n = 8.
Downloaded from ajcn.nutrition.org by guest on February 4, 2015
postprandial glycemia, the area under the glucose concentration of time is that VLDL triacylglycerol synthesis increases in the
by time curve is certainly related to the rate and extent of glucose liver from the metabolism of fructose (50). The absence of a sim-
absorption, but cannot be used unequivocally as an index of either ilar rise in triacylglycerol concentrations in the late postprandial
kinetic variable. Because only half of the sucrose molecule is glu- period after the high-sucrose breakfast may simply have been a
cose and only part of the fructose moiety is converted to glucose factor of time because the peak difference between the diets was
(54), we would expect lower glucose availability from the gut with seen 5.5 h after lunch, whereas the postprandial period in the
a high-sucrose diet and hence a lower area under the curve. In morning was only 4 h. However, it is also possible that the differ-
addition, a significant proportion of the fructose is likely to be ent response to the two meals was a consequence of the different
used for lipogenesis. states in which they were consumed. Breakfast was eaten after an
The dietary differences in blood lipid concentrations can be overnight fast, whereas lunch was consumed only 4 h after a pre-
explained by a combination of the factors described so far and by vious meal. The greater triacylglycerol concentrations after lunch
the unique features of fructose metabolism. The differences in with the high-sucrose diet may have been a result of increased
blood lipid concentrations between the diets together with 99% synthesis, encouraged by the fed state in which the meal was con-
CIs about the zero difference are shown in Figure 9 and Figure 10 sumed. One possible mechanism could be the increased concen-
(a 99% CI was used to reduce the risk of a type I statistical error trations of malonyl-CoA in the fed state, which would promote
resulting from the multiple t testing used to derive the interval). esterification.
The two main troughs, immediately after lunch and after dinner, Whitley et al (55) showed a late rise in triacylglycerol concen-
were a result of the greater peaks in insulin secretion with the trations after a test meal rich in simple carbohydrate compared
high-sucrose diet, thus suppressing mobilization of NEFAs from with a high-fat meal. The same Oxford Lipid Metabolism group
adipose tissue. Three significant peaks are apparent 3–5 h after investigated effects of sequential meals and reported an early rise
each meal for plasma NEFAs, representing higher concentrations in triacylglycerol and NEFA concentrations after lunch, appar-
with the high-sucrose diet than with the high-starch diet. The late ently related to consumption of fat in the breakfast (56). Com-
postprandial increases in NEFA concentrations with the high- pared with the Oxford studies, increases in plasma triacylglycerol
sucrose diet may have been a consequence of the lower insulin in the present study were delayed.
concentrations at these time points, thus facilitating NEFA release Both the raised NEFA and triacylglycerol concentrations with
from adipose tissue. A further explanation could be that there was the high-sucrose diet have a potential role to play in decreasing
simply less available carbohydrate in the high-sucrose diet (as a insulin sensitivity. Studies in rats have shown a close associa-
result of the incomplete conversion of fructose to glucose), neces-
sitating mobilization of fatty acids for oxidation. The nighttime TABLE 5
peak may have been the result of lowered carbohydrate availabil- Mean fasting blood concentrations at the end of each dietary period
ity with the high-sucrose diet, leading to a need for increased lipid
mobilization and oxidation. This is supported by the increased Blood Serum Plasma Serum
Dietary period glucose insulin NEFAs triacylglycerol
glycerol concentrations seen overnight with the high-sucrose diet
compared with the high-starch diet. mmol/L pmol/L mmol/L mmol/L
The clearest difference in triacylglycerol profiles between the High-sucrose diet 4.62 ± 0.31 38.3 ± 13.5 0.25 ± 0.039 0.73 ± 0.16
two diets was seen late in the afternoon, with concentrations rising High-starch diet 4.95 ± 0.28 36.2 ± 19.7 0.22 ± 0.079 0.76 ± 0.18
late after the midday meal with the high-sucrose diet and continu- 1 –
x ± SD. No significant differences were detected with paired t test
ing to rise until the surge of insulin produced by the next meal. comparisons of data (geometrically transformed for fasting insulin
The most plausible explanation in the context of this short period concentrations).1194 DALY ET AL
Downloaded from ajcn.nutrition.org by guest on February 4, 2015
FIGURE 9. Difference between diets (high-sucrose diet 2 high-starch diet) for plasma fatty acid (NEFA) concentrations. Shaded area is the 99%
CI for zero difference between treatments. –x ± SEM; n = 8.
tion between decreased insulin sensitivity with high-sucrose or We also examined the effects of the two diets on glycolytic
high-fructose diets and fasting hypertriglyceridemia (5, 7, 9). Some products. The marked increases in blood lactate and pyruvate
human studies have reported a rise in fasting triacylglycerol con- concentrations after each meal with the high-sucrose but not the
centrations with more moderate intakes of sucrose or fructose, and high-starch diet can be explained most plausibly by the rapid
an association between these increased concentrations and raised metabolism of fructose in the liver.
fasting insulin concentrations may reflect decreased insulin sensi- In conclusion, in young, healthy adults 24-h exposure to a
tivity (57, 58). Although the link between raised fasting and raised sucrose-rich diet compared with a high-starch diet, with all other
postprandial concentrations of triacylglycerol or NEFAs and major dietary components held constant, produced no detectable
insulin resistance is not established, the potential role of raised effect on insulin sensitivity as assessed by a modified insulin tol-
NEFA or triacylglycerol concentrations in decreasing insulin sensi- erance test. However, this lack of effect cannot be extrapolated to
tivity is suspected and has been shown experimentally (59, 60). the population at large. More sedentary, obese persons who are
Furthermore, altered postprandial lipid profiles have been identi- insulin resistant may be more sensitive to alterations in the type
fied as a cardiovascular risk factor (61). of carbohydrate in the diet, a possibility we are addressing in fur-
FIGURE 10. Difference between diets (high-sucrose diet 2 high-starch diet) for plasma triacylglycerol concentrations. Shaded area is the 99% CI
for zero difference between treatments. –x ± SEM; n = 8.INSULIN SENSITIVITY AND DIURNAL METABOLIC PROFILES 1195
ther work. Blood glucose and serum insulin concentrations fol- sitivity in healthy young and old adults. Am J Clin Nutr
lowed the pattern expected for these diets, with increased areas 1990;52:524–8.
under the curve for glucose with the high-starch diet, a probable 15. Chen M, Bergman R, Porte D. Insulin resistance and beta-cell dys-
consequence of the greater availability of glucose with this diet. function in aging: the importance of dietary carbohydrate. J Clin
Endocrinol Metab 1988;67:951–7.
Similarly, the greater rise postprandially in plasma fatty acids
16. Garg A, Grundy S, Unger R. Comparison of effects of high and low
with the high-sucrose diet may have resulted from a combination carbohydrate diets on plasma lipoproteins and insulin sensitivity in
of less available carbohydrate and lower insulin concentrations. patients with mild NIDDM. Diabetes 1992;41:1278–85.
Markedly increased late postprandial triacylglycerol concentra- 17. Howard B, Abbott W, Swinburn B. Evaluation of metabolic effects
tions with the high-sucrose diet may represent increased hepatic of substitution of complex carbohydrates for saturated fat in indi-
synthesis. These last two findings may be involved in the patho- viduals with obesity and NIDDM. Diabetes Care 1991;14:786–95.
genesis of reduced insulin sensitivity, which has been associated 18. Thorburn A, Crapo P, Griver K, Wallace P, Henry R. Long-term
with long-term consumption of such diets. effects of dietary fructose on carbohydrate metabolism in NIDDM.
The detection of such between-diet differences in fit, healthy, Metabolism 1990;39:58–63.
insulin-sensitive persons gives rise to the possibility that such 19. Crapo P, Kolterman O, Henry R. Metabolic consequences of two-
week fructose feeding in diabetic subjects. Diabetes Care
changes would be exaggerated in a more insulin-resistant popu-
1986;9:111–9.
lation. The observation that responses to the experimental diets 20. Turner JL, Bierman EL, Brunzell JD, Chait A. Effect of dietary fruc-
differed with different meals during the day, particularly for tri- tose on triglyceride transport and glucoregulatory hormones in
acylglycerol concentrations, emphasizes the limitations in hypertriglyceridemic men. Am J Clin Nutr 1979;32:1043–50.
assessing the postprandial effects of a nutrient with a single test 21. Dunnigan M, Fyfe T, McKiddie, Crosbie S. The effects of isocaloric
meal. Finally, the key to the differences between these diets may exchange of dietary starch and sucrose on glucose tolerance, plasma
not be relative proportions of disaccharides and polysaccharides, insulin and serum lipids in man. Clin Sci 1970;38:1–9.
Downloaded from ajcn.nutrition.org by guest on February 4, 2015
but rather the presence of fructose within the sucrose diet. 22. Mann J, Truswell A. Effects of isocaloric exchange of dietary
sucrose and starch on fasting serum lipids, postprandial insulin
secretion and alimentary lipaemia in human subjects. Br J Nutr
1972;27:395–405.
REFERENCES 23. Reiser S, Handler HB, Gardner LL, Hallfrisch JG, Michaelis OE
4th, Prather ES. Isocaloric exchange of dietary starch and sucrose in
1. Avogaro P, Crepaldi G. Essential lipemia, obesity and diabetes. Dia- humans. II. Effect on fasting blood insulin, glucose, and glucagon
betologia 1965;1:137 (abstr). and on insulin and glucose response to a sucrose load. Am J Clin
2. Reaven GM. The role of insulin resistance in human disease. Diabetes Nutr 1979;32:2206–16.
1988;37:1595–607. 24. Reiser S, Bohn E, Hallfrisch J, Michaelis OE 4th, Keeney M,
3. Bonadonna R, Groop L, Kraemer N, Ferrannini E, Del Prato S,
Prather ES. Serum insulin and glucose in hyperinsulinemic subjects
DeFronzo R. Effect of insulin on oxidative and non-oxidative path-
fed three different levels of sucrose. Am J Clin Nutr
ways of free fatty acid metabolism in human obesity. Am J Physiol
1981;34:2348–58.
1992;263:E79–84.
25. Beck-Nielsen H, Pedersen O, Lindskov HO. Impaired cellular
4. Mikines K, Sonne B, Farrell P, Tronier B, Galbo H. Effect of physical
insulin binding and insulin sensitivity induced by high-fructose
exercise on sensitivity and responsiveness to insulin in humans. Am J
feeding in normal subjects. Am J Clin Nutr 1980;33:273–8.
Physiol 1988;254:E248–59.
26. Koivisto V, Yki-Jarvinen H. Fructose and insulin sensitivity in
5. Himsworth H. The dietetic factor determining the glucose tolerance
patients with type 2 diabetes. J Intern Med 1993;233:145–53.
and sensitivity to insulin of normal men. Clin Sci 1935;2:68–94.
27. Kiens B, Richter EA. Types of carbohydrate in an ordinary diet
6. Pagliosotti M, Shahrokhi K, Moscarello M. Involvement of liver and
skeletal muscle in sucrose-induced insulin resistance: dose response affect insulin action and muscle substrates in humans. Am J Clin
studies. Am J Physiol 1994;266:R1637–44. Nutr 1996;63:47–53.
7. Storlien LH, Kraegen EW, Jenkins AB, Chisholm DJ. Effect of sucrose 28. Coulston AM, Hollenbeck CB, Donner CC, Williams R, Chiou YM,
vs starch diets on in vivo insulin action, thermogenesis, and obesity in Reaven GM. Metabolic effects of added dietary sucrose in individ-
rats. Am J Clin Nutr 1988;47:420–7. uals with non-insulin dependent diabetes mellitus. Metabolism
8. Gutman R, Basilico M, Bernal C, Chicco A, Lombardo Y. Long-term 1985;10:962–6.
hypertriglyceridemia and glucose intolerance in rats fed chronically an 29. Crapo PA, Kolterman OG. Metabolic consequences of 2-week fruc-
isocaloric sucrose-rich diet. Metabolism 1987;36:1013–20. tose feeding in normal subjects. Am J Clin Nutr 1984;39:525–34.
9. Wright DW, Hansen RI, Mondon CE, Reaven GM. Sucrose-induced 30. Garrow JS, James WPT, eds. Human nutrition and dietetics. 9th ed.
insulin resistance in the rat: modulation by exercise and diet. Am J London: Churchill Livingstone, 1993.
Clin Nutr 1983;38:879–83. 31. Tonolo G, Ciccarese M, Brizzi P, et al. Cyclical variation of plasma
10. Thorburn AW, Storlien LH, Jenkins AB, Khouri S, Kraegen EE. Fruc- lipids, apolipoproteins, and lipoprotein (a) during the menstrual
tose-induced in vivo insulin resistance and elevated plasma triglyc- cycle of normal women. Am J Physiol 1995;269:E1101–5.
eride levels in rats. Am J Clin Nutr 1989;49:1155–63. 32. Schijf CPT, van der Mooren MJ, Doesburg WH, Thomas CMG, Rol-
11. Zavoroni I, Sander S, Scott S, Reaven G. Effect of fructose feeding on land R. Differences in serum lipids, lipoproteins, sex hormone bind-
insulin secretion and insulin action in the rat. Metabolism ing globulin and testosterone between the follicular and the luteal
1980;29:970–3. phase of the menstrual cycle. Acta Endocrinol 1993;129:130–3.
12. Storlien LH, Oakes N, Pan D, Kusonoki M, Jenkins A. Syndromes of 33. Siri WS. The gross composition of the body. In: Lawrence TH,
insulin resistance in the rat: inducement by diet and amelioration with Tobias CA, eds. Advances in biological and medical physics. New
benfluorex. Diabetes 1993;42:457–62. York: Academic Press, 1956:239–80.
13. Hulman S, Falkner B. The effect of excess dietary sucrose on growth, 34. Durnin JVGA, Womersley. Body fat assessed from total body den-
blood pressure, and metabolism in developing Sprague-Dawley rats. sity and its estimation from skinfold thickness: measurements on
Pediatr Res 1994;36:95–101. 481 men and women aged from 16–72 years. Br J Nutr
14. Fukagawa NK, Anderson JJ, Hageman G, Young VR, Minaker KL. 1974;32:77–97.
High-carbohydrate, high-fiber diets increase peripheral insulin sen- 35. Akinmokun A, Selby PL, Ramaiya K, Alberti KGMM. The short1196 DALY ET AL
insulin tolerance test for the determination of insulin sensitivity—a 1994;17:1453–9.
comparison with the euglycaemic clamp. Diabet Med 1992;9:432–7. 49. Nestler JE, Barlascini CO, Clore JN, Blackard WG. Absorption
36. Harrison J, Hodson AW, Skillen AW, Stappenbeck R. Blood glucose, characteristic of breakfast determines insulin sensitivity and carbo-
lactate, pyruvate, glycerol, hydroxybutyrate and aceto-acetate meas- hydrate tolerance for lunch. Diabetes Care 1988;11:755–60.
urements in man using a centrifugal attachment with a fluorimetric 50. Mayes PA. Intermediary metabolism of fructose. Am J Clin Nutr
attachment. J Clin Chem Clin Biochem 1988;26:141–6. 1993;58(suppl):754S–65S.
37. Soden P, Hopkins C. Microdiet. Lancashire, United Kingdom: 51. Brundin T, Wahren J. Whole body and splanchnic oxygen con-
Department of Mathematics and Computer Science, University of sumption and blood flow after oral ingestion of fructose or glucose.
Salford, 1985. Am J Physiol 1993;264:E504–13.
38. Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgate
52. McGarry JD, Takabayashi Y, Foster DW. The role of malonyl CoA
DAT. McCance and Widdowson’s the composition of foods. 5th ed.
in the co-ordination of fatty acid synthesis and oxidation in isolated
London: Her Majesty’s Stationery Office, 1991.
rat hepatocytes. J Biol Chem 1978;253:9294–300.
39. Schofield WN, Schofield C, James WPT. Basal metabolic rate-review
53. Topping DL, Mayes PA. The immediate effects of insulin and fruc-
and prediction. Hum Nutr Clin Nutr 1985;39(suppl):1–96.
40. Department of Health. Dietary reference values for food energy and tose on the metabolism of the perfused liver. Changes in lipopro-
nutrients for the United Kingdom. London: Her Majesty’s Stationery tein secretion, fatty acid oxidation and esterification, lipogenesis
Office, 1991. (Reports on Public Health and Medical Subjects, no. and carbohydrate metabolism. Biochem J 1972;126:295–311.
41.) 54. Björkman O, Felig P. Role of the kidney in the metabolism of fruc-
41. Mullen A, Luke V. Photographic food atlas. An atlas of average food tose in 60hr fasted humans. Diabetes 1982;31:516–20.
portions. Liverpool, United Kingdom: University of Liverpool 55. Whitley HA, Humphreys SM, Samra JS, et al. Metabolic responses
Department of Medicine and The Royal Liverpool University Hospi- to isoenergetic meals containing different proportions of carbohy-
tal Trust Department of Medicine, 1995. drate and fat. Br J Nutr 1997;78:15–26.
Downloaded from ajcn.nutrition.org by guest on February 4, 2015
42. Bingham S, Cummings JH. The use of 4-aminobenzoic acid as a 56. Fielding BA, Callow J, Owen RM, Samra JS, Matthews DR, Frayn
marker to validate the completeness of 24 hr urine collections in man. KN. Postprandial lipemia: the origin of an early peak studied by
Clin Sci (Colch) 1983;64:629–35. specific dietary fatty acid intake during sequential meals. Am J
43. Crapo PA, Reaven G, Olefsky J. Plasma glucose and insulin responses Clin Nutr 1996;63:36–41.
to orally administered simple and complex carbohydrates. Diabetes 57. Reiser S, Bickard MC, Hallfrisch J, Michaelis OE, Prather EJ.
1976;25:741–7. Blood lipids and their distribution in lipoprotein in hyperinsuline-
44. Wolever TM, Jenkins DJ, Kalmusky J, et al. Glycaemic response to mic subjects fed three different levels of sucrose. J Nutr
pasta: effect of surface area, degree of cooking and protein enrich- 1981;111:1045–57.
ment. Diabetes Care 1986;9:401–4. 58. Hallfrisch J, Reiser S, Prather ES. Blood lipid distribution of
45. Morgan LM, Tredger JA, Wright J, Marks V. The effect of soluble and
hyperinsulinemic men consuming three levels of fructose. Am J
insoluble-fibre supplementation on post-prandial glucose tolerance,
Clin Nutr 1983;37:740–8.
insulin and gastric inhibitory polypeptide secretion in healthy sub-
59. Laville M, Rigalleau V, Riou JP, Beylot M. Respective role of
jects. Br J Nutr 1990;64:103–10.
plasma non-esterified fatty acid oxidation and total lipid oxidation
46. d’Emden MC, Marwick TH, Dreghorn J, Howlett VL, Cameron DP.
Post-prandial glucose and insulin responses to different types of in lipid-induced insulin resistance. Metabolism 1995;44:639–44.
spaghetti and bread. Diabetes Res Clin Pract 1987;3:221–6. 60. Boden G, Chen X. Effects of fat on glucose uptake and utilization
47. Donatelli M, Russo V, Bucalo ML, Scarpinato A. Plasma alanine and in patients with non-insulin dependent diabetes mellitus. J Clin
lactate concentrations following glucose ingestion in normal and Invest 1995;96:1261–8.
NIDDM subjects. Diabetes Res 1992;20:121–6. 61. Hughes TA, Elam MB, Applegate WB, et al. Post-prandial lipopro-
48. Ercan N, Gannon MC, Nuttal FQ. Effect of added fat on the plasma tein responses in hypertriglyceridemic individuals with and without
glucose and insulin response to ingested potato given in various cardiovascular disease. Metab Clin Exp 1995;44:1082–98.
combinations as two meals in normal individuals. Diabetes CareYou can also read
