- Email: [email protected]
The effect of high carbohydrate diet on glucose tolerance in patients with type 2 diabetes mellitus
Diabetes Research and Clinical Practice 57 (2002) 163– 170 www.elsevier.com/locate/diabres
The effect of high carbohydrate diet on glucose tolerance in patients with type 2 diabetes mellitus Nobuyuki Komiyama a, Takashi Kaneko b, Akio Sato b, Wataru Sato c, Kaoru Asami d, Toshimasa Onaya a, Masato Tawata a,* a
Third Department of Internal Medicine, Yamanashi Medical Uni6ersity, Shimokato 1110, Tamaho, Yamanashi 409 -3898, Japan Department of En6ironmental Health, Yamanashi Medical Uni6ersity, Shimokato 1110, Tamaho, Yamanashi 409 -3898, Japan c Department of Medical Information, Yamanashi Medical Uni6ersity, Shimokato 1110, Tamaho, Yamanashi 409 -3898, Japan d Nutritional Di6ision, Yamanashi Medical Uni6ersity, Shimokato 1110, Tamaho, Yamanashi 409 -3898, Japan
b
Received 11 October 2001; received in revised form 18 February 2002; accepted 13 March 2002
Abstract The effect of high carbohydrate (hc) diet on glucose tolerance and on lipid profiles in patients with type 2 diabetes mellitus is contradicted. Japanese patients with mild type 2 diabetes mellitus were allocated either 55% standard carbohydrate (sc) or 80% high carbohydrate diets for 1 week, and OGTT and lipid profiles were examined. Then the diet was crossed over for another week, and OGTT and other identical parameters were re-evaluated. High carbohydrate diet improved the area under the glucose concentration– time curve (AUG) in 16/24 patients, and significantly increased and decreased 1,5-anhydroglucitol and homeostasis model assessment insulin resistance (HOMA-R) as a whole, respectively. Fasting plasma glucose (FPG) hc/sc ratio was inversely correlated with HOMA-R on a standard carbohydrate diet. High carbohydrate diet significantly decreased LDL- and HDL-cholesterol, whereas it significantly increased triglyceride. Furthermore, hc/sc ratios of the lipid parameters were inversely correlated with the respective parameters on standard carbohydrate diet. The present study indicates that high carbohydrate diet improved glucose tolerance depending on patients and the improvement in FPG was predicted by HOMA-R on a standard carbohydrate diet. The effect of high carbohydrate diet on glucose tolerance and lipid profiles should be investigated through a long-term study in the future. © 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: High carbohydrate diet; Glucose tolerance; HOMA-R; Insulinogenic index; LDL-cholesterol; HDL-cholesterol; Triglyceride
This study was supported by Grants of Yamanashi Medical University and UTY Science Foundation. * Corresponding author. Tel.: + 81-55-273-9597; fax: + 8155-273-9685 E-mail address: [email protected] (M. Tawata).
1. Introduction The optimal calorie for patients with type 2 diabetes mellitus is individualized based on the standard body weight (SBW), physical activity, etc. On the other hand, recommended carbohy-
0168-8227/02/$ - see front matter © 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 8 - 8 2 2 7 ( 0 2 ) 0 0 0 5 3 - 0
164
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
drate composition ranges about 55– 60% of the total calorie in the United State of America [1] as well as in Japan. However, there is no evidence that the present carbohydrate composition is optimal to all patients with type 2 diabetes mellitus. Himsworth showed that a low carbohydrate diet impaired glucose tolerance in normal subjects while a high carbohydrate diet improved it [2]. However, Wilkerson et al. negated the effect of carbohydrate restriction on glucose tolerance [3]. Then the beneficial effect of a high carbohydrate diet on glucose tolerance was re-evaluated in normal subjects [4,5]. Recently, it was confirmed that even a single low carbohydrate meal significantly impaired glucose tolerance [6], and that the impairment of glucose tolerance was related with a decreased insulinogenic index in healthy control subjects [6,7]. On the other hand, Brunzell et al. reported that an 85% high carbohydrate diet improved glucose tolerance in patients with mild diabetes mellitus [8]. However, there are also contradicting reports [9 – 14] concerning the effect of a high carbohydrate diet on glucose tolerance in patients with diabetes mellitus. Therefore, we do not know yet whether a high carbohydrate diet really improves or impairs glucose tolerance in patients with diabetes mellitus. It is also possible that a high carbohydrate diet improves glucose tolerance in some and impairs it in others depending on the patients with diabetes mellitus. As a first step to elucidate which one of these is more likely, we undertook a pilot study to observe the effects of a high carbohydrate diet on glucose tolerance and serum lipid profiles in Japanese patients with mild type 2 diabetes mellitus.
2. Methods
2.1. Subjects The present study was undertaken according to the principles of the Declaration of Helsinki, and the study was approved by the Institutional Review Board of Yamanashi Medical University. Full informed consent was obtained from all participants before the study.
Patients with type 2 diabetes mellitus, who were hospitalized at the metabolic ward of Yamanashi Medical University Hospital, were recruited for the study. Those who were taking any medications for glycemic control were excluded from the study. In the end, 24 patients participated in the present study. The background characteristics of all these participants are described in Table 1. Their mean (9 SD) age was 54.59 11.0 years old, body-mass index (BMI) was 26.39 3.4 (kg/m2), and HbA1c was 7.29 1.1%.
2.2. Study design The SBW was calculated by [height (m)]2 × 22, and the total calorie was determined in the range of 24 –26 kcal/kg SBW. After attaining less than 150 mg/dl of FPG at least for 1 week, patients were allocated either a standard carbohydrate diet (sc: 55–60% carbohydrate, 15–18% protein and 22–30% fat) or a high carbohydrate diet (hc: 78– 80% carbohydrate diet, 14–16% protein and 4–8% fat) on iso-caloric condition. After 7 days either on the standard or high carbohydrate diet, 75 g OGTT and other parameters including serum lipids were examined on the following day of fasting. Then, the diet was crossed over for another 7 days, and the second OGTT and other identical parameters were re-evaluated. The order of the standard carbohydrate diet or the high carbohydrate diet was determined based on the date (odd or even) of admission of each subject. Those who were allocated high carbohydrate diet and standard carbohydrate diet first were designated as groups H and S, respectively. Complex carbohydrates were used as a source of carbohydrate. The ratio of saturated (S), monounsaturated (M) and polyunsaturated (P) fatty acids was 3:4:3 in both diets. Parameters were abbreviated as the following: FPG for fasting plasma glucose; PG1 for 1-h plasma glucose during OGTT; PG2 for 2-h plasma glucose during OGTT; AUG for area under the glucose concentration–time curve during OGTT; IIND for insulinogenic index (IRI at 30 min during OGTT-fasting IRI)/(plasma glucose at 30 min during OGTT-FPG) [15]; AUI for area under the insulin concentration–time curve
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
165
(Linco Research, St. Louis, MO). Serum 1,5-anhydroglucitol was measured by an enzymatic method using pyranose oxidase after pretreated by glucokinase. HDL- and LDL-cholesterol were measured by using cholesterol esterase and cholesterol oxidase. Serum triglyceride concentration was measured by using lipoprotein lipase, glycerol kinase and glycerol 3%-phosphate oxidase.
during OGTT; HDL-cholesterol for serum high density lipoprotein-cholesterol; LDL-cholesterol for serum low density lipoprotein-cholesterol; BMI for body-mass index (body weight/height2 [kg/m2]); HOMA-R for homeostasis model assessment insulin resistance [16] (FPG× fasting IRI/ 405); and HOMA-B for homeostasis model assessment b-cell function (360× fasting IRI/ [FPG-63]) [16]. The suffixes of sc and hc indicate the parameters on standard and high carbohydrate diets, respectively. Furthermore, the suffix of hc/sc indicates the ratio of the data of the parameters on the high carbohydrate diet divided by the data on the standard carbohydrate diet. Plasma glucose was measured by a glucose oxidation method, and serum insulin was measured by a two-site immunoradiometric assay. Serum leptin was measured by radioimmunoassay
2.3. Statistical analysis All continuous variables are expressed as the mean9 SD. The differences in the background characteristics or various parameters between the groups H and S were analyzed by Mann–Whitney U-test or by Fisher’s exact test. The effect of the stratification of various parameters on glucose tolerance was evaluated by Fisher’s PLSD test.
Table 1 Background characteristics of the patients Patients
Age
Sex
BMI
HbA1c (%)
Group
Period (days)
AUG hc/sc
AUI hc/sc
IIND hc/sc
c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 c12 c13 c14 c15 c16 c17 c18 c19 c20 c21 c22 c23 c24
69 37 44 65 52 53 41 45 62 57 61 57 41 41 63 72 30 67 54 63 59 58 53 64
F M M M M F M M F M M M M F M F M F F M M F F M
28.13 27.28 22.74 24.65 25.43 24.37 26.63 36.22 33.49 25.44 23.15 26.38 21.95 28.90 21.30 25.90 29.08 27.73 23.95 23.55 24.35 25.37 29.65 26.85
6.2 6.6 7.5 7.4 9.7 9.5 8.7 6.4 7.8 6.3 7.9 8.1 7.8 5.9 7.3 6.4 6.3 5.3 5.9 8.2 6.5 7.6 6.9 7.3
H S S H S H S S H S H S S S H H H S H H S H S S
56 17 11 23 17 19 24 12 20 22 36 18 31 115 20 31 16 20 14 14 34 44 37 25
0.996 1.006 0.935 1.008 1.068 0.944 1.054 0.878 1.114 0.895 0.950 0.951 0.542 0.972 0.927 1.106 0.940 0.957 1.250 0.998 0.830 0.908 0.835 1.070
1.185 1.090 1.108 0.926 0.771 0.991 1.281 1.279 0.633 1.403 1.061 0.806 0.827 0.776 0.823 0.924 1.151 0.930 1.098 0.507 0.405 0.308 0.831 0.599
1.505 1.927 2.574 0.874 0.428 3.687 0.937 2.750 0.714 1.677 1.317 0.582 3.728 0.921 1.165 1.359 0.531 1.221 0.906 2.648 1.414 0.569 1.021 2.105
M, male; F, female; BMI, body mass index; Group, first-allocated diet; S, standard carbohydrate diet; H, high carbohydrate diet; Period, the period of hospitalization before the study; AUG, area under the glucose concentration–time curve during OGTT; AUI, area under the insulin concentration–time curve during OGTT; IIND, insulinogenic index; hc/sc ratio, the ratio of the data under high hycarbohydrate diet to the data under standard carbohydrate diet.
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
166
Table 2 Contingency tables of the improvement rates in glucose tolerance depending on the age (A) and the period of hospitalization (B) A
Yonger age (n= 13)
P value
Improved
Worsened
Group H Group S
2 7
1 3
B
Shorter hospitalization (n= 12)
Group H Group S
Improved
Worsened
4 4
2 2
Older age (n = 11)
P value
Improved
Worsened
n.s.
5 2
3 1
P value
Longer hospitalization (n = 12)
n.s.
Improved
Worsened
3 5
2 2
n.s.
P value
n.s.
Improved and Worsened indicate that glucose tolerance was improved (AUG hc/sc ratio B1) and worsened (AUG hc/sc ratio \1), respectively, by high carbohydrate diet. Groups H and S indicate those whose first-allocated diet were high carbohydrate and standard carbohydrate diets, respectively. n.s.,: not significant by Fisher’s exact test. The figures in the table indicate the real number of patients. (A) The patients were stratified into two groups depending on the age. Since there were two patients with 57 years old, so each group has either 11 or 13 patients. Statistical analysis also revealed no difference when we compared 11 patients in younger age and 13 patients in older age. (B) The patients were stratified into two groups depending on the length of hospitalization before the study.
The Wilcoxon signed-ranks test was applied to evaluate statistical significance between each parameter of the standard carbohydrate diet versus the high carbohydrate diet. Linear regression analysis was performed and Pearson’s correlation coefficient was obtained. P value below 0.05 was considered statistically significant.
3. Results Table 1 shows the background characteristics and some parameters of the participants. The comparison between the groups H and S revealed that the former group showed significantly higher age (59.19 11.2 versus 50.699.8 years old), FPG, PG1 and AUG (865.4 9 108.2 versus 746.99 115.0 mg/dl/2 h) than the latter group. There was no significant difference in other background characteristics or in other parameters between the two groups. However, the improvement rates of glucose tolerance (AUG hc/sc ratio B 1) and insulin response (IIND hc/sc ratio \ 1) were almost the same in both groups (7/11 and 6/11 cases, respectively, in group H, and 9/13 and 9/13 cases, respectively, in group S) (Table 1). There was no significant correlation between AUG hc/sc
ratio and age (r=0.285, P = 0.1794), FPG, PG1 or AUG (r= 0.165, P =0.4467). Furthermore, there was no difference in the improvement rate of glucose tolerance by high carbohydrate diet between the groups H and S even when the patients were stratified by the age (Table 2A). There was no significant correlation between AUG hc/sc ratio and HbA1c (r= −0.017, P= 0.9382), BMI (r= 0.147, P=0.4986) or the period of hospitalization (r= − 0.106, P= 0.6265). There was no significant correlation between the period of hospitalization and hc/sc ratios of FPG, PG1 or PG2 either. Furthermore, there was no difference in the improvement rate of glucose tolerance by high carbohydrate diet between the groups H and S even when the patients were stratified by the period of hospitalization (Table 2B). However, when the patients were stratified into two groups by the period of hospitalization, those with longer hospitalization showed significantly lower FPG hc/sc ratio (0.929 0.08 versus 1.039 0.9) than those with shorter hospitalization. Regarding the data on standard carbohydrate diet, we observed significant correlation between AUG and PG1, FPG or PG2 (Table 3). On the contrary, there was a significant inverse correla-
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
tion between AUG and IIND or HOMA-B (Table 3). In addition, there was also a significant inverse correlation between FPG and HOMA-B or IIND (Table 3). There was a significant correlation between AUI and HOMA-B, IIND or HOMA-R (Table 3). Serum 1,5-anhydroglucitol showed a significant correlation with BMI or HOMA-B (Table 3). There was also a significant correlation between serum leptin and BMI or HOMA-R (Table 3). Moreover, we observed a significant increase in 1,5-anhydroglucitol (Fig. 1A), and a significant decrease in HOMA-R (Fig. 1B) as a whole by the high carbohydrate diet. We also observed a significant correlation between the AUG hc/sc ratio and PG1 hc/sc (r = 0.852, PB0.0001), PG2 hc/sc (r = 0.802, PB 0.0001) or FPG hc/sc (r=0.513, P = 0.0094) ratios. Furthermore, there was a significant inverse correlation between the AUG hc/sc ratio and HOMA-B hc/sc (r = − 0.492, P = 0.0159) or IIND hc/sc (r = − 0.471, P =0.0191) ratios. Table 3 Correlations between various parameters on standard carbohydrate diet Parameters AUG AUG AUG AUG AUG FPG FPG AUI AUI AUI AG AG Leptin Leptin
PG1 FPG PG2 IIND HOMA-B HOMA-B IIND HOMA-B IIND HOMA-R HOMA-B BMI BMI HOMA-R
r
P value
0.929 0.891 0.750 −0.592 −0.480 −0.658 −0.643 0.779 0.653 0.571 0.559 0.555 0.692 0.502
B0.0001 B0.0001 B0.0001 0.0023 0.0193 0.0004 0.0006 B0.0001 0.0005 0.0037 0.0048 0.0042 0.0010 0.0390
Correlations were analyzed between various parameters on standard carbohydrate diet. AUG, area under the glucose concentration–time curve during OGTT; FPG, fasting plasma glucose; PG1, 1-h plasma glucose during OGTT; PG2, 2-h plasma glucose during OGTT; IIND, insulinogenic index; HOMA-B, homeostasis model assessment b-cell function; AUI, area under the insulin concentration–time curve during OGTT; HOMA-R, homeostasis model assessment insulin resistance; AG, l,5-anhydroglucitol; BMI, body-mass index.
167
There was also a significant correlation between the HOMA-B hc/sc and HOMA-R hc/sc (r= 0.526, P=0.0090) ratios. Furthermore, we observed a significant inverse correlation between the FPG hc/sc ratio and HOMA-R on the standard carbohydrate diet (Fig. 2). Concerning lipid profiles, there were significant reductions in LDL- and HDL-cholesterol, and a significant increase in triglyceride by the high carbohydrate diet (Fig. 3A). Furthermore, the hc/sc ratio of each lipid parameter was inversely correlated with the respective parameter on the standard carbohydrate diet (Fig. 3B).
4. Discussion Since we adopted a quasi-random method concerning the order of study diets, we observed significant differences in age and the degree of glucose intolerance between groups H and S. However, based on the various data shown in the results, we think that the order of study diet per se did not substantially affect the efficacy of high carbohydrate diet to undermine the present study despite these differences. The data on the standard carbohydrate diet shown in Table 3 including serum leptin [17] can be expected and seems to be all reasonable. The high carbohydrate diet decreased HOMAR, which is considered to be an ‘insulin resistance’ index [16], as a whole. Furthermore, the IIND hc/sc ratio was increased in 15/24 patients by the high carbohydrate diet, and the AUG hc/sc ratio showed an inverse correlation with the IIND hc/ sc ratio. Therefore, we speculate that the high carbohydrate diet improved glucose tolerance by increasing insulin secretion from pancreatic b-cells in addition to the decrease in ‘insulin resistance’. The increase in glucose effectiveness and insulin secretion by a high carbohydrate diet had been shown [4,5]. Brunzell et al. reported an improvement in glucose tolerance by a high carbohydrate diet [8]. By calculating HOMA-R using their data [8], we found a decrease in HOMA-R by the high carbohydrate diet in all subjects except one case. Chen et al. reported an increase in the insulin
168
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
Fig. 1. Changes in 1,5-anhydroglucitol (1,5-AG) (A) and HOMA-R (B) between standard carbohydrate (sc) and high carbohydrate (hc) diets. Values were expressed as the mean 9 SD. An increase in 1,5-AG and a decrease in HOMA-R by a high carbohydrate diet were statistically significant by the Wilcoxon signed-ranks test. P value less than 0.05 was considered statistically significant. HOMA-R, homeostasis model assessment insulin resistance.
sensitivity index and insulin secretion by a high carbohydrate diet [18]. We also reported an increase in the insulinogenic index along with the improvement in glucose tolerance by a high carbohydrate diet in healthy control subjects [6,7]. Thus, the present study agrees with the previous studies in these respects. However, since we observed decreased insulin secretion from pancreatic b-cells in nine patients by the high carbohydrate diet in the present study, insulin response to a high carbohydrate diet also seems to be different depending on patients. There are contradictory reports concerning the effect of a high carbohydrate diet on glucose tolerance [9–14]. Although the high carbohydrate diet significantly increased 1,5-anhydroglucitol as a whole and improved glucose tolerance in twothirds of patients in the present study, it also impaired glucose tolerance in others. Therefore, the present study suggests that a high carbohydrate diet may improve glucose tolerance in some and impair it in others depending on the patients. It is expected that FGP of the patients who are hospitalized for longer period are likely to be improved by a high carbohydrate diet irrespective
of the order of the test diets. The present study also predicts that FGP of those with HOMA-R greater than 1.61 on a standard carbohydrate diet is expected to be improved by a high carbohydrate diet. Since these predictions are based on a relatively small number of patients and a rela-
Fig. 2. Correlation between FPG hc/sc ratio and HOMA-R on standard carbohydrate (sc) diet. The ordinate indicates the FPG hc/sc ratio and the abscissa indicates HOMA-R sc. Linear regression analysis was performed and Pearson’s correlation coefficient was obtained. P value less than 0.05 was considered statistically significant. FPG, fasting plasma glucose; HOMA-R, homeostasis model assessment insulin resistance.
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
169
Fig. 3. Changes in lipid parameters between standard carbohydrate (sc) and high carbohydrate (hc) diets (A), and correlation between hc/sc ratios of lipid parameters and respective parameters on standard carbohydrate diet (B). (A) The Wilcoxon signed-ranks test revealed that a high carbohydrate diet significantly decreased LDL-cholesterol and HDL-cholesterol, while it significantly increased TG. (B) The hc/sc ratio was obtained by calculating the value on high carbohydrate divided by the value on a standard carbohydrate diet. The ordinates indicate hc/sc ratios of lipid parameters and the abscissas indicate the respective lipid parameters on a standard carbohydrate diet. Linear regression analysis was performed and Pearson’s correlation coefficient was obtained. P value less than 0.05 was considered statistically significant. LDL-cholesterol, low density lipoprotein-cholesterol; HDL-cholesterol, high density lipoprotein-cholesterol; TG, triglyceride.
tively short study period, we do not know yet to what extent the prediction can be extrapolated to clinical settings. However, the present study seems to indicate that the response to a high carbohydrate diet may be different depending on patients and the application of a high carbohydrate diet in clinical settings may have to be considered individually. A better parameter for the prediction of improvement in glucose tolerance by a high carbohydrate diet needs to be pursued by a longterm study in the future, which would eventually
allow us to establish isolation of the protein (P), fat (F) and carbohydrate (C) ratio for each patient. Concerning the effect on lipid profiles, the high carbohydrate diet significantly decreased LDLand HDL-cholesterol and increased triglyceride, which agrees to the previous reports [4,5,19]. To our surprise, the higher the lipid parameters on a standard carbohydrate diet, the greater the decrease one displays in these values by a high carbohydrate diet. Therefore, high triglyceride on
170
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
a standard carbohydrate diet per se may not be a contraindication of a high carbohydrate diet. In conclusion, the present study shows that an 80% high carbohydrate diet for 1 week reduced HOMA-R in patients with mild type 2 diabetes mellitus and improved glucose tolerance in two thirds of the patients. Moreover, improvement in FPG was predicted from HOMA-R on a standard carbohydrate diet. Hypertriglyceridemia did not seem to be a contraindication of a high carbohydrate diet at least to Japanese patients with mild type 2 diabetes mellitus. It is certain that the effect of a high carbohydrate diet on glucose tolerance and lipid profiles in Japanese patients with type 2 diabetes mellitus should be investigated through a long-term study in the future.
Acknowledgements We express our sincere gratitude to Keiko Ichinose and Sachiko Takei for their excellent assistance for the preparation of this manuscript.
References [1] American Diabetes Association, Clinical practice recommendations 2000, Diabetes Care 23 (Suppl.) (2000) S43 – S46. [2] H.P. Himsworth, The dietetic factor determining the glucose tolerance and sensitivity to insulin of healthy men, Clin. Sci. 2 (1935) 67 –94. [3] H.L.C. Wilkerson, C. Hyman, M. Kaufman, A.C. McCuistion, J.O. Francis, Diagnostic evaluation of oral glucose tolerance tests in nondiabetic subjects after various levels of carbohydrate intake, N. Engl. J. Med. 262 (1960) 1047 – 1053. [4] B.V. Howard, W.G. Abbott, B.A. Swinburn, Evaluation of metabolic effects of substitution of complex carbohydrates for saturated fat in individuals with obesity and NIDDM, Diabetes Care 14 (1991) 786 –795. [5] B.A. Swinburn, V.L. Boyce, R.N. Bergman, B.V. Howard, C. Bogardus, Deterioration in carbohydrate metabolism and lipoprotein changes induced by modern, high fat diet in Pima Indians and Caucasians, J. Clin. Endocrinol. Metab. 73 (1991) 156 – 165. [6] T. Kaneko, P.-Y. Wang, M. Tawata, A. Sato, Low carbohydrate intake before oral glucose-tolerance tests,
Lancet 352 (1998) 289. [7] P.-Y. Wang, T. Kaneko, Y. Wang, M. Tawata, A. Sato, Impairment of glucose tolerance in normal adults following lowered carbohydrate intake, Tohoku J. Exp. Med. 189 (1999) 59 – 70. [8] J.D. Brunzell, R.L. Lerner, W.R. Hazzard, D. Porte Jr., E.L. Bierman, Improved glucose tolerance with high carbohydrate feeding in mild diabetes, N. Engl. J. Med. 284 (1971) 521 – 524. [9] A. Garg, A. Bonanome, S.M. Grundy, Z.J. Zhang, R.H. Unger, Comparison of high-carbohydrate diet with highmonounsaturated-fat diet in patients with non-insulin-dependent diabetes mellitus, N. Engl. J. Med. 319 (1988) 829 – 834. [10] M. Parillo, A.A. Rivellese, A.V. Ciardullo, et al., Highmonounsaturated-fat/low-carbohydrate diet improves peripheral insulin sensitivity in non-insulin-dependent diabetic patients, Metab., Clin. Exp. 41 (1992) 1373 – 1378. [11] A. Garg, S.M. Grundy, M. Koffler, Effect of high carbohydrate intake on hyperglycemia, islet function, and plasma lipoproteins in NIDDM, Diabetes Care 15 (1992) 1572 – 1580. [12] A. Garg, J.P. Bantle, R.R. Henry, A.M. Coulston, K.A. Griver, S.K. Raatz, Effects of varying carbohydrate content of diet in patients with non-insulin-dependent diabetes mellitus, J. Am. Med. Assoc. 271 (1994) 1421 – 1428. [13] M. Parillo, R. Giacco, A.V. Ciardullo, A.A. Rivellese, G. Riccardi, Does high carbohydrate diet have different effects in NIDDM patients treated with diet alone or hypoglycemic drugs?, Diabetes Care 19 (1996) 498 – 500. [14] M.C. Gannon, F.Q. Nuttall, S.A. Westphal, S. Fang, N. Ercan-Fang, Acute metabolic response to high-carbohydrate, high-starch meals compared with moderate-carbohydrate, low-starch meals in subjects with type 2 diabetes, Diabetes Care 21 (1998) 1619 – 1926. [15] A. Ohneda, K. Watanabe, Y. Maruhama, H. Itabashi, K. Horigome, Changes in glucose tolerance and insulin response in subclinical diabetes after treatment, Tohoku J. Exp. Med. 122 (1977) 199 – 207. [16] D.R. Matthews, J.P. Hosker, A.S. Rudenski, B.A. Naylor, D.F. Treacher, R.C. Turner, Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man, Diabetologia 28 (1985) 412 – 419. [17] R.V. Considine, M.K. Sinha, M.L. Heiman, et al., Serum immunoreactive-leptin concentrations in normal-weight and obese humans, N. Engl. J. Med. 334 (1996) 292 – 295. [18] M. Chen, R.N. Bergman, D. Porte Jr., Insulin resistance and b-cell dysfunction in aging: the importance of dietary carbohydrate, J. Clin. Endocrinol. Metab. 67 (1988) 951 – 957. [19] S.M. Grundy, Comparison of monounsaturated fatty acids and carbohydrate for lowering plasma cholesterol, N. Engl. J. Med. 314 (1986) 745 – 748.
The effect of high carbohydrate diet on glucose tolerance in patients with type 2 diabetes mellitus Nobuyuki Komiyama a, Takashi Kaneko b, Akio Sato b, Wataru Sato c, Kaoru Asami d, Toshimasa Onaya a, Masato Tawata a,* a
Third Department of Internal Medicine, Yamanashi Medical Uni6ersity, Shimokato 1110, Tamaho, Yamanashi 409 -3898, Japan Department of En6ironmental Health, Yamanashi Medical Uni6ersity, Shimokato 1110, Tamaho, Yamanashi 409 -3898, Japan c Department of Medical Information, Yamanashi Medical Uni6ersity, Shimokato 1110, Tamaho, Yamanashi 409 -3898, Japan d Nutritional Di6ision, Yamanashi Medical Uni6ersity, Shimokato 1110, Tamaho, Yamanashi 409 -3898, Japan
b
Received 11 October 2001; received in revised form 18 February 2002; accepted 13 March 2002
Abstract The effect of high carbohydrate (hc) diet on glucose tolerance and on lipid profiles in patients with type 2 diabetes mellitus is contradicted. Japanese patients with mild type 2 diabetes mellitus were allocated either 55% standard carbohydrate (sc) or 80% high carbohydrate diets for 1 week, and OGTT and lipid profiles were examined. Then the diet was crossed over for another week, and OGTT and other identical parameters were re-evaluated. High carbohydrate diet improved the area under the glucose concentration– time curve (AUG) in 16/24 patients, and significantly increased and decreased 1,5-anhydroglucitol and homeostasis model assessment insulin resistance (HOMA-R) as a whole, respectively. Fasting plasma glucose (FPG) hc/sc ratio was inversely correlated with HOMA-R on a standard carbohydrate diet. High carbohydrate diet significantly decreased LDL- and HDL-cholesterol, whereas it significantly increased triglyceride. Furthermore, hc/sc ratios of the lipid parameters were inversely correlated with the respective parameters on standard carbohydrate diet. The present study indicates that high carbohydrate diet improved glucose tolerance depending on patients and the improvement in FPG was predicted by HOMA-R on a standard carbohydrate diet. The effect of high carbohydrate diet on glucose tolerance and lipid profiles should be investigated through a long-term study in the future. © 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: High carbohydrate diet; Glucose tolerance; HOMA-R; Insulinogenic index; LDL-cholesterol; HDL-cholesterol; Triglyceride
This study was supported by Grants of Yamanashi Medical University and UTY Science Foundation. * Corresponding author. Tel.: + 81-55-273-9597; fax: + 8155-273-9685 E-mail address: [email protected] (M. Tawata).
1. Introduction The optimal calorie for patients with type 2 diabetes mellitus is individualized based on the standard body weight (SBW), physical activity, etc. On the other hand, recommended carbohy-
0168-8227/02/$ - see front matter © 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 8 - 8 2 2 7 ( 0 2 ) 0 0 0 5 3 - 0
164
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
drate composition ranges about 55– 60% of the total calorie in the United State of America [1] as well as in Japan. However, there is no evidence that the present carbohydrate composition is optimal to all patients with type 2 diabetes mellitus. Himsworth showed that a low carbohydrate diet impaired glucose tolerance in normal subjects while a high carbohydrate diet improved it [2]. However, Wilkerson et al. negated the effect of carbohydrate restriction on glucose tolerance [3]. Then the beneficial effect of a high carbohydrate diet on glucose tolerance was re-evaluated in normal subjects [4,5]. Recently, it was confirmed that even a single low carbohydrate meal significantly impaired glucose tolerance [6], and that the impairment of glucose tolerance was related with a decreased insulinogenic index in healthy control subjects [6,7]. On the other hand, Brunzell et al. reported that an 85% high carbohydrate diet improved glucose tolerance in patients with mild diabetes mellitus [8]. However, there are also contradicting reports [9 – 14] concerning the effect of a high carbohydrate diet on glucose tolerance in patients with diabetes mellitus. Therefore, we do not know yet whether a high carbohydrate diet really improves or impairs glucose tolerance in patients with diabetes mellitus. It is also possible that a high carbohydrate diet improves glucose tolerance in some and impairs it in others depending on the patients with diabetes mellitus. As a first step to elucidate which one of these is more likely, we undertook a pilot study to observe the effects of a high carbohydrate diet on glucose tolerance and serum lipid profiles in Japanese patients with mild type 2 diabetes mellitus.
2. Methods
2.1. Subjects The present study was undertaken according to the principles of the Declaration of Helsinki, and the study was approved by the Institutional Review Board of Yamanashi Medical University. Full informed consent was obtained from all participants before the study.
Patients with type 2 diabetes mellitus, who were hospitalized at the metabolic ward of Yamanashi Medical University Hospital, were recruited for the study. Those who were taking any medications for glycemic control were excluded from the study. In the end, 24 patients participated in the present study. The background characteristics of all these participants are described in Table 1. Their mean (9 SD) age was 54.59 11.0 years old, body-mass index (BMI) was 26.39 3.4 (kg/m2), and HbA1c was 7.29 1.1%.
2.2. Study design The SBW was calculated by [height (m)]2 × 22, and the total calorie was determined in the range of 24 –26 kcal/kg SBW. After attaining less than 150 mg/dl of FPG at least for 1 week, patients were allocated either a standard carbohydrate diet (sc: 55–60% carbohydrate, 15–18% protein and 22–30% fat) or a high carbohydrate diet (hc: 78– 80% carbohydrate diet, 14–16% protein and 4–8% fat) on iso-caloric condition. After 7 days either on the standard or high carbohydrate diet, 75 g OGTT and other parameters including serum lipids were examined on the following day of fasting. Then, the diet was crossed over for another 7 days, and the second OGTT and other identical parameters were re-evaluated. The order of the standard carbohydrate diet or the high carbohydrate diet was determined based on the date (odd or even) of admission of each subject. Those who were allocated high carbohydrate diet and standard carbohydrate diet first were designated as groups H and S, respectively. Complex carbohydrates were used as a source of carbohydrate. The ratio of saturated (S), monounsaturated (M) and polyunsaturated (P) fatty acids was 3:4:3 in both diets. Parameters were abbreviated as the following: FPG for fasting plasma glucose; PG1 for 1-h plasma glucose during OGTT; PG2 for 2-h plasma glucose during OGTT; AUG for area under the glucose concentration–time curve during OGTT; IIND for insulinogenic index (IRI at 30 min during OGTT-fasting IRI)/(plasma glucose at 30 min during OGTT-FPG) [15]; AUI for area under the insulin concentration–time curve
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
165
(Linco Research, St. Louis, MO). Serum 1,5-anhydroglucitol was measured by an enzymatic method using pyranose oxidase after pretreated by glucokinase. HDL- and LDL-cholesterol were measured by using cholesterol esterase and cholesterol oxidase. Serum triglyceride concentration was measured by using lipoprotein lipase, glycerol kinase and glycerol 3%-phosphate oxidase.
during OGTT; HDL-cholesterol for serum high density lipoprotein-cholesterol; LDL-cholesterol for serum low density lipoprotein-cholesterol; BMI for body-mass index (body weight/height2 [kg/m2]); HOMA-R for homeostasis model assessment insulin resistance [16] (FPG× fasting IRI/ 405); and HOMA-B for homeostasis model assessment b-cell function (360× fasting IRI/ [FPG-63]) [16]. The suffixes of sc and hc indicate the parameters on standard and high carbohydrate diets, respectively. Furthermore, the suffix of hc/sc indicates the ratio of the data of the parameters on the high carbohydrate diet divided by the data on the standard carbohydrate diet. Plasma glucose was measured by a glucose oxidation method, and serum insulin was measured by a two-site immunoradiometric assay. Serum leptin was measured by radioimmunoassay
2.3. Statistical analysis All continuous variables are expressed as the mean9 SD. The differences in the background characteristics or various parameters between the groups H and S were analyzed by Mann–Whitney U-test or by Fisher’s exact test. The effect of the stratification of various parameters on glucose tolerance was evaluated by Fisher’s PLSD test.
Table 1 Background characteristics of the patients Patients
Age
Sex
BMI
HbA1c (%)
Group
Period (days)
AUG hc/sc
AUI hc/sc
IIND hc/sc
c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 c12 c13 c14 c15 c16 c17 c18 c19 c20 c21 c22 c23 c24
69 37 44 65 52 53 41 45 62 57 61 57 41 41 63 72 30 67 54 63 59 58 53 64
F M M M M F M M F M M M M F M F M F F M M F F M
28.13 27.28 22.74 24.65 25.43 24.37 26.63 36.22 33.49 25.44 23.15 26.38 21.95 28.90 21.30 25.90 29.08 27.73 23.95 23.55 24.35 25.37 29.65 26.85
6.2 6.6 7.5 7.4 9.7 9.5 8.7 6.4 7.8 6.3 7.9 8.1 7.8 5.9 7.3 6.4 6.3 5.3 5.9 8.2 6.5 7.6 6.9 7.3
H S S H S H S S H S H S S S H H H S H H S H S S
56 17 11 23 17 19 24 12 20 22 36 18 31 115 20 31 16 20 14 14 34 44 37 25
0.996 1.006 0.935 1.008 1.068 0.944 1.054 0.878 1.114 0.895 0.950 0.951 0.542 0.972 0.927 1.106 0.940 0.957 1.250 0.998 0.830 0.908 0.835 1.070
1.185 1.090 1.108 0.926 0.771 0.991 1.281 1.279 0.633 1.403 1.061 0.806 0.827 0.776 0.823 0.924 1.151 0.930 1.098 0.507 0.405 0.308 0.831 0.599
1.505 1.927 2.574 0.874 0.428 3.687 0.937 2.750 0.714 1.677 1.317 0.582 3.728 0.921 1.165 1.359 0.531 1.221 0.906 2.648 1.414 0.569 1.021 2.105
M, male; F, female; BMI, body mass index; Group, first-allocated diet; S, standard carbohydrate diet; H, high carbohydrate diet; Period, the period of hospitalization before the study; AUG, area under the glucose concentration–time curve during OGTT; AUI, area under the insulin concentration–time curve during OGTT; IIND, insulinogenic index; hc/sc ratio, the ratio of the data under high hycarbohydrate diet to the data under standard carbohydrate diet.
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
166
Table 2 Contingency tables of the improvement rates in glucose tolerance depending on the age (A) and the period of hospitalization (B) A
Yonger age (n= 13)
P value
Improved
Worsened
Group H Group S
2 7
1 3
B
Shorter hospitalization (n= 12)
Group H Group S
Improved
Worsened
4 4
2 2
Older age (n = 11)
P value
Improved
Worsened
n.s.
5 2
3 1
P value
Longer hospitalization (n = 12)
n.s.
Improved
Worsened
3 5
2 2
n.s.
P value
n.s.
Improved and Worsened indicate that glucose tolerance was improved (AUG hc/sc ratio B1) and worsened (AUG hc/sc ratio \1), respectively, by high carbohydrate diet. Groups H and S indicate those whose first-allocated diet were high carbohydrate and standard carbohydrate diets, respectively. n.s.,: not significant by Fisher’s exact test. The figures in the table indicate the real number of patients. (A) The patients were stratified into two groups depending on the age. Since there were two patients with 57 years old, so each group has either 11 or 13 patients. Statistical analysis also revealed no difference when we compared 11 patients in younger age and 13 patients in older age. (B) The patients were stratified into two groups depending on the length of hospitalization before the study.
The Wilcoxon signed-ranks test was applied to evaluate statistical significance between each parameter of the standard carbohydrate diet versus the high carbohydrate diet. Linear regression analysis was performed and Pearson’s correlation coefficient was obtained. P value below 0.05 was considered statistically significant.
3. Results Table 1 shows the background characteristics and some parameters of the participants. The comparison between the groups H and S revealed that the former group showed significantly higher age (59.19 11.2 versus 50.699.8 years old), FPG, PG1 and AUG (865.4 9 108.2 versus 746.99 115.0 mg/dl/2 h) than the latter group. There was no significant difference in other background characteristics or in other parameters between the two groups. However, the improvement rates of glucose tolerance (AUG hc/sc ratio B 1) and insulin response (IIND hc/sc ratio \ 1) were almost the same in both groups (7/11 and 6/11 cases, respectively, in group H, and 9/13 and 9/13 cases, respectively, in group S) (Table 1). There was no significant correlation between AUG hc/sc
ratio and age (r=0.285, P = 0.1794), FPG, PG1 or AUG (r= 0.165, P =0.4467). Furthermore, there was no difference in the improvement rate of glucose tolerance by high carbohydrate diet between the groups H and S even when the patients were stratified by the age (Table 2A). There was no significant correlation between AUG hc/sc ratio and HbA1c (r= −0.017, P= 0.9382), BMI (r= 0.147, P=0.4986) or the period of hospitalization (r= − 0.106, P= 0.6265). There was no significant correlation between the period of hospitalization and hc/sc ratios of FPG, PG1 or PG2 either. Furthermore, there was no difference in the improvement rate of glucose tolerance by high carbohydrate diet between the groups H and S even when the patients were stratified by the period of hospitalization (Table 2B). However, when the patients were stratified into two groups by the period of hospitalization, those with longer hospitalization showed significantly lower FPG hc/sc ratio (0.929 0.08 versus 1.039 0.9) than those with shorter hospitalization. Regarding the data on standard carbohydrate diet, we observed significant correlation between AUG and PG1, FPG or PG2 (Table 3). On the contrary, there was a significant inverse correla-
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
tion between AUG and IIND or HOMA-B (Table 3). In addition, there was also a significant inverse correlation between FPG and HOMA-B or IIND (Table 3). There was a significant correlation between AUI and HOMA-B, IIND or HOMA-R (Table 3). Serum 1,5-anhydroglucitol showed a significant correlation with BMI or HOMA-B (Table 3). There was also a significant correlation between serum leptin and BMI or HOMA-R (Table 3). Moreover, we observed a significant increase in 1,5-anhydroglucitol (Fig. 1A), and a significant decrease in HOMA-R (Fig. 1B) as a whole by the high carbohydrate diet. We also observed a significant correlation between the AUG hc/sc ratio and PG1 hc/sc (r = 0.852, PB0.0001), PG2 hc/sc (r = 0.802, PB 0.0001) or FPG hc/sc (r=0.513, P = 0.0094) ratios. Furthermore, there was a significant inverse correlation between the AUG hc/sc ratio and HOMA-B hc/sc (r = − 0.492, P = 0.0159) or IIND hc/sc (r = − 0.471, P =0.0191) ratios. Table 3 Correlations between various parameters on standard carbohydrate diet Parameters AUG AUG AUG AUG AUG FPG FPG AUI AUI AUI AG AG Leptin Leptin
PG1 FPG PG2 IIND HOMA-B HOMA-B IIND HOMA-B IIND HOMA-R HOMA-B BMI BMI HOMA-R
r
P value
0.929 0.891 0.750 −0.592 −0.480 −0.658 −0.643 0.779 0.653 0.571 0.559 0.555 0.692 0.502
B0.0001 B0.0001 B0.0001 0.0023 0.0193 0.0004 0.0006 B0.0001 0.0005 0.0037 0.0048 0.0042 0.0010 0.0390
Correlations were analyzed between various parameters on standard carbohydrate diet. AUG, area under the glucose concentration–time curve during OGTT; FPG, fasting plasma glucose; PG1, 1-h plasma glucose during OGTT; PG2, 2-h plasma glucose during OGTT; IIND, insulinogenic index; HOMA-B, homeostasis model assessment b-cell function; AUI, area under the insulin concentration–time curve during OGTT; HOMA-R, homeostasis model assessment insulin resistance; AG, l,5-anhydroglucitol; BMI, body-mass index.
167
There was also a significant correlation between the HOMA-B hc/sc and HOMA-R hc/sc (r= 0.526, P=0.0090) ratios. Furthermore, we observed a significant inverse correlation between the FPG hc/sc ratio and HOMA-R on the standard carbohydrate diet (Fig. 2). Concerning lipid profiles, there were significant reductions in LDL- and HDL-cholesterol, and a significant increase in triglyceride by the high carbohydrate diet (Fig. 3A). Furthermore, the hc/sc ratio of each lipid parameter was inversely correlated with the respective parameter on the standard carbohydrate diet (Fig. 3B).
4. Discussion Since we adopted a quasi-random method concerning the order of study diets, we observed significant differences in age and the degree of glucose intolerance between groups H and S. However, based on the various data shown in the results, we think that the order of study diet per se did not substantially affect the efficacy of high carbohydrate diet to undermine the present study despite these differences. The data on the standard carbohydrate diet shown in Table 3 including serum leptin [17] can be expected and seems to be all reasonable. The high carbohydrate diet decreased HOMAR, which is considered to be an ‘insulin resistance’ index [16], as a whole. Furthermore, the IIND hc/sc ratio was increased in 15/24 patients by the high carbohydrate diet, and the AUG hc/sc ratio showed an inverse correlation with the IIND hc/ sc ratio. Therefore, we speculate that the high carbohydrate diet improved glucose tolerance by increasing insulin secretion from pancreatic b-cells in addition to the decrease in ‘insulin resistance’. The increase in glucose effectiveness and insulin secretion by a high carbohydrate diet had been shown [4,5]. Brunzell et al. reported an improvement in glucose tolerance by a high carbohydrate diet [8]. By calculating HOMA-R using their data [8], we found a decrease in HOMA-R by the high carbohydrate diet in all subjects except one case. Chen et al. reported an increase in the insulin
168
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
Fig. 1. Changes in 1,5-anhydroglucitol (1,5-AG) (A) and HOMA-R (B) between standard carbohydrate (sc) and high carbohydrate (hc) diets. Values were expressed as the mean 9 SD. An increase in 1,5-AG and a decrease in HOMA-R by a high carbohydrate diet were statistically significant by the Wilcoxon signed-ranks test. P value less than 0.05 was considered statistically significant. HOMA-R, homeostasis model assessment insulin resistance.
sensitivity index and insulin secretion by a high carbohydrate diet [18]. We also reported an increase in the insulinogenic index along with the improvement in glucose tolerance by a high carbohydrate diet in healthy control subjects [6,7]. Thus, the present study agrees with the previous studies in these respects. However, since we observed decreased insulin secretion from pancreatic b-cells in nine patients by the high carbohydrate diet in the present study, insulin response to a high carbohydrate diet also seems to be different depending on patients. There are contradictory reports concerning the effect of a high carbohydrate diet on glucose tolerance [9–14]. Although the high carbohydrate diet significantly increased 1,5-anhydroglucitol as a whole and improved glucose tolerance in twothirds of patients in the present study, it also impaired glucose tolerance in others. Therefore, the present study suggests that a high carbohydrate diet may improve glucose tolerance in some and impair it in others depending on the patients. It is expected that FGP of the patients who are hospitalized for longer period are likely to be improved by a high carbohydrate diet irrespective
of the order of the test diets. The present study also predicts that FGP of those with HOMA-R greater than 1.61 on a standard carbohydrate diet is expected to be improved by a high carbohydrate diet. Since these predictions are based on a relatively small number of patients and a rela-
Fig. 2. Correlation between FPG hc/sc ratio and HOMA-R on standard carbohydrate (sc) diet. The ordinate indicates the FPG hc/sc ratio and the abscissa indicates HOMA-R sc. Linear regression analysis was performed and Pearson’s correlation coefficient was obtained. P value less than 0.05 was considered statistically significant. FPG, fasting plasma glucose; HOMA-R, homeostasis model assessment insulin resistance.
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
169
Fig. 3. Changes in lipid parameters between standard carbohydrate (sc) and high carbohydrate (hc) diets (A), and correlation between hc/sc ratios of lipid parameters and respective parameters on standard carbohydrate diet (B). (A) The Wilcoxon signed-ranks test revealed that a high carbohydrate diet significantly decreased LDL-cholesterol and HDL-cholesterol, while it significantly increased TG. (B) The hc/sc ratio was obtained by calculating the value on high carbohydrate divided by the value on a standard carbohydrate diet. The ordinates indicate hc/sc ratios of lipid parameters and the abscissas indicate the respective lipid parameters on a standard carbohydrate diet. Linear regression analysis was performed and Pearson’s correlation coefficient was obtained. P value less than 0.05 was considered statistically significant. LDL-cholesterol, low density lipoprotein-cholesterol; HDL-cholesterol, high density lipoprotein-cholesterol; TG, triglyceride.
tively short study period, we do not know yet to what extent the prediction can be extrapolated to clinical settings. However, the present study seems to indicate that the response to a high carbohydrate diet may be different depending on patients and the application of a high carbohydrate diet in clinical settings may have to be considered individually. A better parameter for the prediction of improvement in glucose tolerance by a high carbohydrate diet needs to be pursued by a longterm study in the future, which would eventually
allow us to establish isolation of the protein (P), fat (F) and carbohydrate (C) ratio for each patient. Concerning the effect on lipid profiles, the high carbohydrate diet significantly decreased LDLand HDL-cholesterol and increased triglyceride, which agrees to the previous reports [4,5,19]. To our surprise, the higher the lipid parameters on a standard carbohydrate diet, the greater the decrease one displays in these values by a high carbohydrate diet. Therefore, high triglyceride on
170
N. Komiyama et al. / Diabetes Research and Clinical Practice 57 (2002) 163–170
a standard carbohydrate diet per se may not be a contraindication of a high carbohydrate diet. In conclusion, the present study shows that an 80% high carbohydrate diet for 1 week reduced HOMA-R in patients with mild type 2 diabetes mellitus and improved glucose tolerance in two thirds of the patients. Moreover, improvement in FPG was predicted from HOMA-R on a standard carbohydrate diet. Hypertriglyceridemia did not seem to be a contraindication of a high carbohydrate diet at least to Japanese patients with mild type 2 diabetes mellitus. It is certain that the effect of a high carbohydrate diet on glucose tolerance and lipid profiles in Japanese patients with type 2 diabetes mellitus should be investigated through a long-term study in the future.
Acknowledgements We express our sincere gratitude to Keiko Ichinose and Sachiko Takei for their excellent assistance for the preparation of this manuscript.
References [1] American Diabetes Association, Clinical practice recommendations 2000, Diabetes Care 23 (Suppl.) (2000) S43 – S46. [2] H.P. Himsworth, The dietetic factor determining the glucose tolerance and sensitivity to insulin of healthy men, Clin. Sci. 2 (1935) 67 –94. [3] H.L.C. Wilkerson, C. Hyman, M. Kaufman, A.C. McCuistion, J.O. Francis, Diagnostic evaluation of oral glucose tolerance tests in nondiabetic subjects after various levels of carbohydrate intake, N. Engl. J. Med. 262 (1960) 1047 – 1053. [4] B.V. Howard, W.G. Abbott, B.A. Swinburn, Evaluation of metabolic effects of substitution of complex carbohydrates for saturated fat in individuals with obesity and NIDDM, Diabetes Care 14 (1991) 786 –795. [5] B.A. Swinburn, V.L. Boyce, R.N. Bergman, B.V. Howard, C. Bogardus, Deterioration in carbohydrate metabolism and lipoprotein changes induced by modern, high fat diet in Pima Indians and Caucasians, J. Clin. Endocrinol. Metab. 73 (1991) 156 – 165. [6] T. Kaneko, P.-Y. Wang, M. Tawata, A. Sato, Low carbohydrate intake before oral glucose-tolerance tests,
Lancet 352 (1998) 289. [7] P.-Y. Wang, T. Kaneko, Y. Wang, M. Tawata, A. Sato, Impairment of glucose tolerance in normal adults following lowered carbohydrate intake, Tohoku J. Exp. Med. 189 (1999) 59 – 70. [8] J.D. Brunzell, R.L. Lerner, W.R. Hazzard, D. Porte Jr., E.L. Bierman, Improved glucose tolerance with high carbohydrate feeding in mild diabetes, N. Engl. J. Med. 284 (1971) 521 – 524. [9] A. Garg, A. Bonanome, S.M. Grundy, Z.J. Zhang, R.H. Unger, Comparison of high-carbohydrate diet with highmonounsaturated-fat diet in patients with non-insulin-dependent diabetes mellitus, N. Engl. J. Med. 319 (1988) 829 – 834. [10] M. Parillo, A.A. Rivellese, A.V. Ciardullo, et al., Highmonounsaturated-fat/low-carbohydrate diet improves peripheral insulin sensitivity in non-insulin-dependent diabetic patients, Metab., Clin. Exp. 41 (1992) 1373 – 1378. [11] A. Garg, S.M. Grundy, M. Koffler, Effect of high carbohydrate intake on hyperglycemia, islet function, and plasma lipoproteins in NIDDM, Diabetes Care 15 (1992) 1572 – 1580. [12] A. Garg, J.P. Bantle, R.R. Henry, A.M. Coulston, K.A. Griver, S.K. Raatz, Effects of varying carbohydrate content of diet in patients with non-insulin-dependent diabetes mellitus, J. Am. Med. Assoc. 271 (1994) 1421 – 1428. [13] M. Parillo, R. Giacco, A.V. Ciardullo, A.A. Rivellese, G. Riccardi, Does high carbohydrate diet have different effects in NIDDM patients treated with diet alone or hypoglycemic drugs?, Diabetes Care 19 (1996) 498 – 500. [14] M.C. Gannon, F.Q. Nuttall, S.A. Westphal, S. Fang, N. Ercan-Fang, Acute metabolic response to high-carbohydrate, high-starch meals compared with moderate-carbohydrate, low-starch meals in subjects with type 2 diabetes, Diabetes Care 21 (1998) 1619 – 1926. [15] A. Ohneda, K. Watanabe, Y. Maruhama, H. Itabashi, K. Horigome, Changes in glucose tolerance and insulin response in subclinical diabetes after treatment, Tohoku J. Exp. Med. 122 (1977) 199 – 207. [16] D.R. Matthews, J.P. Hosker, A.S. Rudenski, B.A. Naylor, D.F. Treacher, R.C. Turner, Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man, Diabetologia 28 (1985) 412 – 419. [17] R.V. Considine, M.K. Sinha, M.L. Heiman, et al., Serum immunoreactive-leptin concentrations in normal-weight and obese humans, N. Engl. J. Med. 334 (1996) 292 – 295. [18] M. Chen, R.N. Bergman, D. Porte Jr., Insulin resistance and b-cell dysfunction in aging: the importance of dietary carbohydrate, J. Clin. Endocrinol. Metab. 67 (1988) 951 – 957. [19] S.M. Grundy, Comparison of monounsaturated fatty acids and carbohydrate for lowering plasma cholesterol, N. Engl. J. Med. 314 (1986) 745 – 748.