TABLE 6.
Intervention trials of breakfast cereals and diabetes1
| Authors (reference) | Quality rating | Subjects and study location | Study design | Diet | Outcomes | Key results |
| Diabetic subjects | ||||||
| Colagiuri et al. (127) | Positive | 8 subjects (males and females, aged 42–69 y) with NIDDM but treated with insulin; Australia | Randomized crossover design study. Subjects given 3 test breakfast meals in random order of 3 d, 48 h apart. Meals prepared by a dietitian; blood collected at 30 min intervals for 3 h. | 3 meals matched in energy (470–480 kJ) and CHO (54–60 g) | Postprandial glucose (PG), serum insulin (SI), and C-peptide (CP) responses | Postprandial responses to meals A and B were similar. Meal C (muesli and skim milk) produced lower PG and SI responses to both other meals (P < 0.02) and less endogenous insulin secretion CP: C = 62.8 vs. B = 99.8 vs. A = 89.8 pmol ⋅ min#x22121 ⋅ mL#x22121 (P < 0.05). |
| A: 2 eggs, 2 slices whole-meal toast, orange juice | ||||||
| B: 2 Weet-Bix, whole-meal toast, | ||||||
| C: 75 g muesli | ||||||
| All with milk, tea, or coffee | ||||||
| Golay et al. (126) | Positive | 14 adults with type 2 NIDDM, treated with insulin; mean age = 69 y; mean BMI (in kg/m2) = 29; Switzerland | Randomized crossover design study: two 1-wk periods separated by 1-wk washout. Patients followed normal diet but provided with SRS or FRS breakfast (46 g CHO) and morning snack food | 2 breakfast meals: | Plasma glucose and insulin throughout the day, and over 3 h after oral GTT at end of diet period | Mean daily blood glucose concentration 21% lower with muesli breakfast compared with cornflakes (7.7 ± 0.7 vs. 9.7 ± 0.9; P = 0.023) |
| SRS: 65 g muesli + 120 mL whole milk | Fasting insulin was 17% lower after the muesli compared with cornflakes (126 vs. 150; P < 0.05). No change in Hb A1c. | |||||
| FRS: 35 g cornflakes + 12 g sugar + 120 mL whole milk | Patients reduced daily insulin doses from mean of 28 ± 4 to 24 ± 4 U (P < 0.05) during the muesli period. | |||||
| “Switching at breakfast only from standard cereals to slow-release starch cereals improves the CHO metabolism of diabetic patients” (p135) | ||||||
| Tappy et al. (129) | Positive | 8 subjects (males and females, aged 49–65 y) with NIDDM; Switzerland | Randomized crossover design study. Subjects given 4 breakfast meals, in random order on different days, each providing 35 g CHO. | 3 meals with milk and an extruded breakfast cereal with 4, 6, or 8.4 g β-glucan from oats, and 1 with whole-meal bread, ham, and jam | AUC of postprandial plasma glucose and insulin measured over 4 h | Compared with the noncereal breakfast, all 3 cereal breakfasts significantly decreased the peak and average glucose and insulin increments. Maximum increases in plasma glucose after cereal breakfast were 67% (P < 0.05), 42%, and 38% (P < 0.001) with 4, 6, or 8.4 g β-glucan cereal compared with the noncereal breakfast. |
| Wheeler et al. (128) | Positive | 24 subjects (males and females) aged 14–25 y with IDDM; US | Randomized crossover design study. Subjects given 4 test breakfast meals in random order of 4 d, 72 h apart. Meals were prepared by a dietitian; blood collected at 30 min intervals for 3 h. | 4 meals matched for CHO (50 g/1.73 m2 body surface) | AUC over 3 h for: | There was no difference in the AUC for the 2 cereals for plasma glucose. |
| 1) unsweetened cornflake cereal | Plasma glucose | There were no differences in insulin response to any of the meals. | ||||
| 2) sweetened cornflake cereal (40% sucrose) (both Kellogg Co) | Free insulin | “Equivalent gram amounts of CHO as presweetened cereals are not detrimental to people with IDDM compared to unsweetened cereals.” (p458) | ||||
| 3) sucrose alone | ||||||
| 4) glucose alone | ||||||
| All consumed with water | ||||||
| Tsihlias et al. (131) | Positive | 72 subjects with type 2 NIDDM; mean age = 62 y, mean BMI = 27; Canada | Randomized parallel design with 3 treatment arms each lasting 6 mo: 10% energy from high-GI breakfast cereal, low-GI breakfast cereal, or high-MUFA intake with no breakfast cereal | High-GI cereals: cornflakes (Nature's Path), puffed rice (Arrowhead Mills), crispy rice (Our Compliments) | Hb A1c | Compared with MUFA treatment, with cereal breakfasts subjects consumed 10% more energy from CHO. |
| Low-GI cereals: Bran Buds with psyllium (Kellogg) or prototype extruded oat cereal with psyllium | Fasting glucose | There were no significant differences in Hb A1c or fasting glucose between treatments, although cereal treatment groups had higher 8-h plasma insulin concentrations after cereal breakfast compared with the MUFA group (P < 0.05). | ||||
| MUFA-diet subjects given margarine and olive oil | Plasma insulin | “A 10% increase in CHO intake from breakfast cereal had no deleterious effects on glycemic control over 6 mo in subjects with type 2 DM [diabetes mellitus].” (p439) | ||||
| Rendell et al. (136) | Positive | 16 nondiabetic men and women (mean age = 56 y, mean BMI = 30)and18 type 2 NIDDM men and women (mean age = 62 y, mean BMI = 33); US | Randomized crossover design comparing 2 cereal breakfast test meals, and a liquid meal replacement control, each test day separated by 3–7 d washout. Subjects tested after standardized evening meal and overnight. | 65 g oatmeal (7 g fiber) or Prowash barley flakes (23 g fiber) cooked with 360 mL water | AUC postprandial plasma glucose and insulin over 2 h | The AUC glucose and insulin concentrations were significantly lower with the Prowash barley compared with oatmeal in both normal and diabetic subjects. |
| Prowash barely contained 15% β-glucan vs. 5% in oats | In diabetics: | |||||
| Glucose: 83 ± 13 vs. 200 ± 34 mg/dL (P < 0.001) | ||||||
| Insulin: 30 ± 6 vs. 93 ± 32 mIU/dL (P < 0.001) | ||||||
| “Inclusion of foods with CHO that are absorbed slowly is beneficial in both the diabetic and prediabetic states” (p66) | ||||||
| Nondiabetic subjects | ||||||
| Wolever et al. (132) | Positive | 77 healthy nondiabetic men aged 18–75 y; BMI = 18.5–34; divided into normal and high fasting insulin (≥41 pmol/L); Canada | Randomized crossover design study with 2 breakfasts consumed after overnight fast. | 1) High-fiber cereal (Fiber One): 36.8 g available CHO + 36.7 g fiber | Plasma glucose and insulin measured over 2 h after breakfast | In all subjects, 2-h AUC plasma glucose increase was less after high-fiber cereal (107 ± 7 vs .130 ± 8 mmol × min/L; P < 0.001). |
| 2) Low-fiber cereal (Cornflakes): 36.8 g available CHO + 0.8 g dietary fiber) | Insulin peak response was only reduced by high-fiber cereal in hyperinsulinemic men (351 ± 29 vs. 485 ± 55 pmol/L; P = 0.044) but not normal controls. | |||||
| Both consumed with 250 mL low-fat milk. | “RTEC rich in nonviscous cereal fiber reduces glucose responses in normal and hyperinsulinemic men” (p1285) | |||||
| Maki et al. (133) | Positive | 27 healthy nondiabetic nonsmoking men aged 25–54 y with BMI <32; US | Randomized crossover study with two 2-wk treatment periods with a 1-wk washout, incorporating oat- or wheat-based breakfast cereals into usual diet. | Two energy- and fiber-matched cereals provided: | Postprandial glucose and insulin responses over 10 h after a fat-loading breakfast including hot cereal | “Postprandial insulin and glucose responses over 10 h did not differ between [oat and wheat] treatments” (p347) |
| Oat: 76 g/d oat bran RTEC plus 60 g/d hot oatmeal | ||||||
| Wheat: 84 g/d Frosted Mini-Wheats (Kellogg) plus 60 g/d hot rolled-wheat cereal | ||||||
| Hlebowicz et al. (135) | Positive | 12 healthy nondiabetic men and women; mean age = 28 y, mean BMI = 22; Sweden | Randomized crossover design study with 3 breakfast meals taken after an overnight fast, more than 1 wk apart. | Each 50 g cereal served with 300 g sour milk: | 1) GER measured by real-time ultrasonography | The wheat-bran-based cereal resulted in a lower GER (33%) compared with the oat-based cereal (51%) (P < 0.05) but was not significantly different from cornflakes. |
| 1) All-Bran (7.5 g fiber) | 2) Blood glucose from finger-prick samples | There were no significant differences in the 2-h AUC for glucose or satiety. | ||||
| 2) Whole-meal oat flakes (4 g fiber) | 3) Validated satiety score | “Cereal bran flakes slows the GER when compared to oat flakes and corn flakes, probably due to a higher fiber content.” (p1) | ||||
| 3) Cornflakes (1.5 g fiber) | ||||||
| Granfeldt et al. (130) | Positive | 19 healthy nondiabetic mean and women; mean age = 38 y, mean BMI = 22.4; Sweden | Crossover design study with 2 breakfast test meals taken after an overnight fast, 1 wk apart. | 2 oat bran mueslis, made with 3 or 4 g of β-glucan, served with vanilla yogurt, plus a sandwich of white bread, cheese, and butter | AUC of blood glucose and insulin measured over 2 h after breakfast meal | Muesli with 3 g β-glucans gave no significant difference in glycemic response compared with the reference. |
| Meals standardized to a total of 50 g available CHO | Muesli with 4 g β-glucans compared with a reference meal (no muesli) produced a lower glycemic response: | |||||
| Reference meal used muesli with corn rather than oat flakes | AUC: glucose, −29.3%; insulin, −42% (P < 0.05) | |||||
| “4 g β-glucan from oats seems to be a critical level for a significant decreases in glucose and insulin responses in healthy people” (p600) | ||||||
| Kim et al. (134) | Positive | 17 normoglycemic obese women, age 51 y, mean BMI = 33.2; US | Randomized crossover design comparing 5 cereal breakfast test meals, each test day separated by 7-d washout. Subjects tested after standardized evening meal and overnight fast. | 5 energy- and CHO-matched hot cereals (mixed wheat and barley) with 0, 2.5, 5, 7.5, or 10 g β-glucan per serving | Postprandial plasma glucose and insulin measured over 3 h | The 10-g β-glucan cereal significantly reduced peak glucose at 30 min (P < 0.05), but 2-h AUC did not differ between cereals. |
| 10 g β-glucan cereal significantly reduced AUC for insulin (P < 0.05) vs. 0 or 5 g. | ||||||
| “High β-glucan whole-grain products may prove useful in the inclusion of dietary management of hyperglycemia in obese women” (p174) |
1
CHO, carbohydrate; FRS, fast release starch; GER, gastric emptying rate; GI, glycemic index; GTT, glucose tolerance test; Hb A1c, glycated hemoglobin; IDDM, insulin dependent diabetes mellitus; NIDDM, non–insulin-dependent diabetes mellitus; RTEC, ready-to-eat breakfast cereal; SRS, slow release starch.