Garrett Hoyt1, Matthew S. Hickey1, Loren Cordain1, CA1
1Department of Health and Exercise Science, Fort Collins, CO 80523, USA
10 May 2004 - 10 August 2004 - 14 September 2004
In most carbohydrate-containing foods, the blood insulin response is predictable and is closely linked to the food's glycaemic index (GI). A single study, examining whole milk and fermented milk products made from whole milk, recently reported a large dissociation between the GI and insulinaemic index (II) in healthy normal adults. Because the fat component of a food may influence the GI and II, it is unclear if a similar dissociation may exist for skimmed milk in normal adults. We determined the GI and II of both skimmed and whole milk in nine healthy, male (n 6) and female (n 3) subjects (23·6 (sd 1·4) years). No significant (P>0·05) differences existed between GI and II for skimmed and whole milks. Significant (P<0·05) differences were observed between the actual and predicted areas under the insulin curves for both skimmed milk (predicted 1405 (sd 289) pmol×min/l; actual 6152 (sd 1177) pmol×min/l) and whole milk (predicted 1564 (sd 339) pmol×min/l; actual 5939 (sd 1095) pmol×min/l). Consequently, a large and similar dissociation of the GI and II existed for both whole milk (42 (sd 5) and 148 (sd 14)) and skimmed milk (37 (sd 9) and 140 (sd 13)). It is concluded that the dissociation of the GI and II in milk is not related to its fat content.
Glycaemic index, Insulinaemic index, Whole milk, Skimmed milk
AUC, area under the curve, GI, glycaemic index, II, insulinaemic index
The glycaemic index (GI) was developed by Jenkins et al. (1981) to provide additional information to help diabetic patients make appropriate dietary decisions. The GI has important implications for controlling blood glucose concentrations in both diabetic and healthy individuals (Wolever et al. 1995; Björck et al. 2000), but has been criticized for overlooking the contribution of insulin to metabolic disorders (Hollenbeck et al. 1986). More recent studies have illustrated that the glucose response does not always predict the insulin response (Gannon et al. 1988; Holt et al. 1997). By contrasting the predicted and actual insulin response, the relative dissociation of the GI from the insulinaemic index (II) can be determined (Gannon et al. 1986). Dissociations of the insulin response from the glucose response have been demonstrated for oatmeal, kidney beans and lentils in type 2 diabetics (Krezowski et al. 1987) in which glucose was used as the reference food, as well as in beef and fish in healthy subjects (Holt et al. 1997) in which white bread was the reference food. Not all studies of diabetics have been able to show this effect (Aro et al. 1987). In most carbohydrate-containing foods, the blood insulin response is predictable and is closely linked to the food's glucose response (Holt et al. 1997). A single study, examining whole milk and fermented milk products made from whole milk, recently reported a large dissociation between the GI and II for healthy normal subjects in which white bread was used as the reference food (Östman et al. 2001). Because the fat component of a food may influence the GI and II (Collier & O'Dea, 1983), it is unclear if a similar dissociation may exist for skimmed milk in normal subjects.
Nine healthy subjects (six men, mean age 23·3 (sd 1·4) years, mean BMI 23·5 (sd 1·6) kg/m2; three women, mean age 24·3 (sd 1·2) years, mean BMI 24·3 (sd 1·3) kg/m2) were studied on three separate occasions separated by at least 2 d. Only eight subjects were reported for the skimmed milk trials. The subjects recorded their dietary and exercise patterns for 2 d prior to reporting to the laboratory following an 8–10 h fast. The subjects were asked to replicate their dietary and exercise patterns before each subsequent test. Upon arrival at the laboratory a catheter was placed in a forearm vein, and a baseline blood sample was drawn. The catheter was kept patent using normal saline after each blood draw. The subject then consumed one of three test foods (glucose, whole milk and skimmed milk) administered randomly. Each food contained 25 g available carbohydrate. Blood was drawn 15, 30, 45, 60, 90 and 120 min after completion of the meal and was analysed for plasma glucose and insulin concentrations. Glucose was analysed using the hexokinase/glucose-6-phosphate dehydrogenase method, and insulin was measured using the DSL-10-1600 ACTIVE Insulin ELISA Kit (Diagnostic Systems Laboratories, Inc., Webster, TX, USA). Incremental area under the curve (AUC) above baseline was calculated for both glucose and insulin (Wolever et al. 1991). Glucose was set as 100 and the GI and II of whole and skimmed milk were presented as a percentage of the glucose and insulin AUC after 25 g glucose. Prior to data collection the project was approved by the Human Subjects Committee at Colorado State University. Data were analysed using SPSS version 11.5 (Statistical Package for Social Science, SPSS Inc., Chicago, IL, USA). Paired t tests were employed to evaluate dependent variable differences between whole and skimmed milks and α was set at 0·05. The relative dissociation of the GI from the II for both skimmed and whole milks was calculated using a previously described procedure (Gannon et al. 1986). With this procedure the insulin response to a test food is algebraically derived using the glycaemic response to the test food and the glycaemic and insulinaemic response to glucose by the same subjects. The derived value is compared with the actual insulin AUC to show the dissociation of the II from the GI.
The results of the experiment are shown in Table 1 and Fig. 1. Values are reported as means and standard deviations. No significant differences existed between GI and II for skimmed and whole milks. Significant (P<0·05) differences were observed between the actual and predicted areas under the insulin curves for both skimmed milk (predicted 1405 (sd 289) pmol×min/l; actual 6152 (sd 1177) pmol×min/l) and whole milk (predicted 1564 (sd 339) pmol×min/l; actual 5939 (sd 1095) pmol×min/l). Consequently, a large and similar dissociation of the GI and II existed for both whole milk (42 (sd 5) and 148 (sd 14)) and skimmed milk (37 (sd 9) and 140 (sd 13)).
|Whole milk||Skimmed milk|
|Predicted AUC (pmol×min/l)||Actual AUC (pmol×min/l)||t Score||p-Value|
Area under the glucose (A) and insulin (B) curves for glucose, whole and skim milks. (•) glucose; (□) skimmed milk; (○) whole milk.
The novel finding of this experiment was that skimmed milk elicited a disproportionately large insulinaemic response relative to its low glycaemic response in healthy normal subjects. Our results expand upon Gannon and colleagues' work showing skimmed milk to be a potent insulin secretagogue in type 2 diabetic patients (Gannon et al. 1986) and corroborate a previous study demonstrating a similar effect in healthy normal subjects with whole milk and fermented products made from whole milk (Östman et al. 2001).
Östman et al. (2001) established that insulinaemia was greater after their subjects consumed milk products than after an equivalent amount of lactose and water, indicating that some milk component in addition to lactose stimulates insulin secretion. Although these authors suggested that a lipid component in milk may have been responsible, the present results effectively rule out this possibility and suggest that some factor within the protein fraction was responsible for milk's insulinotropic effect. Certain amino acids (tryptophan, leucine, isoleucine and glutamine) are insulinogenic (Schmid et al. 1989). Hence, it has been hypothesized that elevated concentrations of these amino acids in milk may underlie its insulin-stimulating capabilities (Östman et al. 2001). However, the insulinogenic amino acid profile of beef and milk are quite similar (Nutritionist V, Firstdatabank, Indianapolis, IN, USA), and beef has an insulin score of 51 (Holt et al. 1997), whereas the insulin scores for milk products have been reported to range from 89 to 115 (Holt et al. 1997; Östman et al. 2001). Lactose alone has an II of 50 (Östman et al. 2001). Taken together, these data suggest that an additive effect of lactose and amino acids likely underlie milk's insulin-stimulating effects. In support of this notion are data showing that when healthy subjects consumed 50 g protein as lean beef together with 50 g glucose, the insulin response was additive (Krezowski et al. 1986). Nonetheless, we cannot rule out other potential insulin secretagogues in the protein fraction of milk including specific peptides or even endogenous bovine hormones (Koldovský, 1995).
Except for cheese with an insulin score of 45 (Holt et al. 1997) all dairy products (whole milk, skimmed milk, yoghurt, ice cream, cottage cheese and fermented milk products) have been shown to have potent insulinotropic properties that may have far-reaching health effects, given the hypothesis that insulinaemia is a modulator of insulin resistance (Ludwig, 2002). Our data (Fig. 1) confirm the observation of Östman et al. (2001) that consumption of milk induces a reactive hypoglycaemia. Four of our subjects' blood glucose concentrations had dropped below baseline within 60 min after consuming milk. This response is similar to, high glycaemic load carbohydrates, which have been implicated as an underlying cause of certain diseases of insulin resistance (Ludwig, 2002). Even the addition of milk to low GI mixed meals elicits an insulinotropic effect (Liljeberg Elmstahl & Bjorck, 2001). Despite these potentially adverse acute effects, a recent epidemiological report demonstrated dairy consumption to be associated with a lower risk for type 2 diabetes (Pereira et al. 2002). Nevertheless, until well-controlled interventions can corroborate epidemiological associations, our data suggest caution is warranted in recommending higher milk consumption for adults, particularly those at risk for diseases of insulin resistance. Clearly, further research is needed to elucidate milk's potential to influence insulin metabolism adversely.
We would like to thank the Gatorade Sports Science Institute and Colorado State University's Department of Health and Exercise Science for funding this study.
[R1] Aro A Pelkonen R Leino U (1987) Glucose and insulin responses to meals containing milk, lactose, glucose or fructose in subjects with non-insulin-dependent diabetes Diabetes Metab 13 603 – 606
[R2] Björck I Liljeberg H Östman E (2000) Low glycaemic-index foods Br J Nutr 83 Suppl. 1 S149 – S155
[R3] Collier GR O'Dea K (1983) The effect of coingestion of fat on the glucose, insulin and gastric inhibitory polypeptide responses to carbohydrate and protein Am J Clin Nutr 37 941 – 944
[R4] Gannon MC Nuttall FQ Krezowski PA Billington CJ Parker S (1986) The serum insulin and plasma glucose responses to milk and fruit in Type 2 (non-insulin-dependent) diabetic patients Diabetologia 29 784 – 791
[R5] Gannon MC Nuttall FQ Neil BJ Westphal SA (1988) The insulin and glucose responses to meals of glucose plus various proteins in type II diabetic subjects Metabolism 37 1081 – 1088
[R6] Hollenbeck CB Coulston AM Reaven GM (1986) Glycemic effects of carbohydrates: a different perspective Diabetes 9 641 – 647
[R7] Holt S Brand Miller J Petocz P (1997) An insulin index of foods: the insulin demand generated by 1000-kJ portions of common foods Am J Clin Nutr 66 1264 – 1276
[R8] Jenkins D Wolever T Taylor R Barker H Fielden H Baldwin J Bowling A Newman H Jenkins A Goff D (1981) Glycemic index of foods: a physiological basis for carbohydrate exchange Am J Clin Nutr 34 362 – 366
[R9] Koldovský O (1995) Hormones in milk Vitamins and Hormones 77 – 149 Litwack G New York Academic Press
[R18] Krezowski PA Nuttall FQ Gannon MC Bartosh NH (1986) The effect of protein ingestion on the metabolic response to oral glucose in normal individuals Am J Clin Nutr 44 847 – 856
[R10] Krezowski PA Nuttall FQ Gannon MC Billington CJ Parker S (1987) Insulin and glucose responses to various starch-containing foods in type II diabetic subjects Diabetes Care 10 205 – 212
[R11] Liljeberg Elmstahl H Bjorck I (2001) Milk as a supplement to mixed meals may elevate postprandial insulinaemia Eur J Clin Nutr 55 994 – 999
[R12] Ludwig D (2002) The glycemic index: physiological mechanisms relating to obesity, diabetes and cardiovascular disease JAMA 287 2414 – 2423
[R13] Östman E Liljeberg H Björck I (2001) Inconsistency between glycemic and insulinemic responses to regular and fermented milk products Am J Clin Nutr 74 96 – 100
[R14] Pereira MA Jacobs DR van Horn L Slattery ML Kartashov AI Ludwig DS (2002) Dairy consumption, obesity, and the insulin resistance syndrome in young adults: the CARDIA study JAMA 287 2081 – 2089
[R15] Schmid R Schusdziarra V Schulte-Frohlinde E Maier V Classen M (1989) Role of amino acids in stimulation of postprandial insulin, glucagons, and pancreatic polypeptide in humans Pancreas 4 305 – 314
[R16] Wolever T Bentum-Williams A Jenkins D (1995) Physiological modulation of plasma free fatty acid concentrations on diet Diabetes Care 18 962 – 970
[R17] Wolever T Jenkins D Jenkins A Josse R (1991) The glycemic index: methodology and clinical implications Am J Clin Nutr 54 846 – 854