Diets high and low in glycemic index versus high monounsaturated fat diets: effects on glucose and lipid metabolism in NIDDM ND Luscombe 1 , M Noakes 2 * and PM Clifton 2 1 Department of Physiology, University of Adelaide; and 2 CSIRO Division of Human Nutrition, Adelaide, South Australia Objective: To examine the relative effects of high and low glycemic index (GI) carbohydrates, and mono- unsaturated fats on blood glucose and lipid metabolism in NIDDM subjects. Subjects: Fourteen male and seven female variably controlled NIDDM subjects recruited by advertisement. Setting: Free living outpatients. Research design: A repeated measures, within-subject design was used such that each subject consumed three diets: (a) a high-GI diet (53% CHO7 21% fat, 63 GI units (glucose  100)); (b) a low-GI diet (51% CHO 7 23% fat, 43 GI units); and (c) a high-mono high-GI diet (42% CHO735% fat, 59 GI units) in random order and cross-over fashion for four weeks. Approximately 45% energy was provided as key foods which differed in published GI values and speci®cally excluded legumes. Dietary ®bre intake was > 30 g=d on each diet. At the end of each dietary intervention, we measured fasting plasma lipids, glucose, insulin, total glycated plasma protein, fructosamine, LDL and HDL particle size as well as 24 h urinary excretion of glucose and C-peptide. Results: HDL-cholesterol was higher on the low-GI and high-mono high-GI diets compared to the high-GI diet (P < 0.05 for overall diet effect). There were no other signi®cant differences in metabolic control between diets, even when adjusted for BMI, glucose control or gender. Body weight and saturated fat intake remained stable between dietary interventions. Conclusion: High-mono high-GI and high-CHO, low-GI diets are superior to high-CHO, high-GI diets with respect to HDL metabolism but no effect was noted on glucose metabolism in variably controlled NIDDM subjects. Sponsorship: This study was funded by CSIRO Human Nutrition. Descriptors: non-insulin dependent diabetes diet; carbohydrates; glycemic index; monounsaturated fatty acids; plasma lipids Introduction NIDDM is associated with several lipid aberrations, namely, hypertriglyceridemia, elevated VLDL cholesterol and reduced HDL cholesterol and both hyperglycemia and this dyslipidemia are implicated in the development of diabetic complications (De Fronzo et al, 1992; Laakso, 1994). Dietary guidelines for people with NIDDM currently include a diet containing less than 10% from saturated fat, up to 10% energy from polyunsaturated fat, with the distribution of energy from monounsaturated fat and car- bohydrate (CHO) dependent on treatment goals (American Diabetes Association, 1994). The high-CHO diet has been shown to improve glycemic control (Brunzell et al, 1971; Anderson, 1977; Kiehm et al, 1976; Simpson et al, 1979), lower plasma LDL cholesterol concentrations (O'Dea et al, 1989; Riccardi et al, 1984) and reduce insulin requirements (Anderson, 1977; Kiehm et al, 1976). However, some studies have shown that high-CHO diets have hypoglyce- mic and hypolipidemic bene®ts only when they contain a large amount of dietary ®bre (O'Dea et al, 1989; Riccardi et al, 1984) and furthermore, other studies have shown that high-CHO diets low in ®bre result in increased plasma glucose and triacylglycerol, and decreased HDL concentra- tions (Reaven & Olefsky 1974; Brunzell et al, 1974; Ginsberg et al, 1976; Coulston et al, 1983; Sestoft et al, 1985; Coulston et al, 1989) which may persist for more than six weeks (Coulston et al, 1989). It is now well accepted that different foods containing an equal amount of CHO produce a wide range of glycemic responses (Crapo et al, 1977) and this discovery led to the development of the glycemic index (GI) method of ranking CHO-rich foods according to their relative glycemic impact (Jenkins et al, 1981; Foster-Powell & Brand-Miller, 1995). This concept allows the comparison of different foods on the basis of their physiological effects rather than on their chemical composition. A number of studies have shown that, compared to high-GI diets, chronic consumption of low-GI diets similar in ®bre and macronutrient composition, improve glycemic control and plasma lipid pro®les (Miller, 1994). However, the clinical utility of the GI concept continues to be debated. Some studies have failed to show a differ- *Correspondence: Dr M Noakes, PO Box 10041 Gouger Street, ADELAIDE SA 5000, Australia. Contributors: N Luscombe wrote the manuscript and performed all laboratory and some statistical analyses, M Noakes contributed to the study design, and managed the dietary intervention and nutritional analyses and P Clifton contributed to the study design, the manuscript, and performed the statistical analyses. Received 16 June 1998 revised 4 January 1999; accepted 11 January 1999 European Journal of Clinical Nutrition (1999) 53, 473±478 ß 1999 Stockton Press. All rights reserved 0954±3007/99 $12.00 http://www.stockton-press.co.uk/ejcn