8,153
edits
Changes
no edit summary
Glycemic Index (GI) is a measure of how much a food raises blood glucose (blood sugar). A food with a higher Glycemic Index raises the blood sugar more than a food with a lower Glycemic Index. The Glycemic Index value is based on a comparison with a reference food, normally either white bread or glucose. High Glycemic Index foods are linked to a wide variety of health problems, but there is evidence that different people may have widely differing Glycemic Indexes for the same food.
[[File:Glycemic Index Fellrnr.png|none|thumb|500px|A graph of blood glucose (sugar) for high and low glycemic index foods.]]
=Calculating the Glycemic Index=
The Glycemic Index is calculated by measuring blood glucose periodically after the consumption of the food. Typically the measurements are taken just before consuming the food, then at 15, 30, 45, 60, 90, and 120 minutes<ref name="Brand-MillerStockmann2008"/>. These measurements provide a blood glucose curve, and the area between the curve and the baseline measurement taken just before consuming the food is the "incremental Area Under the Curve" (iUAC). The iUAC is compared with the iUAC for a reference food and the value given as a percentage. So a Glycemic Index of 50 means that the food raises the blood sugar by 50% of the reference food. Generally the portion of test food consumed contains 50 grams of carbohydrate. (The iAUC is sometimes called the Post Prandial Glycemic Response or PPGC.)
[[File:Glycemic Index Spike Fellrnr.png|none|thumb|500px| A graph showing a low GI food that has a higher peak blood glucose than a high GI food.]]
Most Glycemic Index examples show two foods, with the high G.I. food having a higher peak blood glucose and a greater incremental Area Under the Curve (iAUC). However, this is not always the case, and Glycemic Index is based only on the iAUC, so a food that rapidly spikes the blood glucose and then the level drops quickly can have a lower iAUC than a food that keeps the blood glucose at a moderate level for longer. This means that Glycemic Index alone may not tell us as much as we'd like.
=Glycemic Index and Health=
High blood glucose levels are linked to many health problems.
* IGT is also linked with metabolic syndrome, which is the cluster of health problems including obesity, high blood pressure, non-alcoholic fatty liver disease, elevated triglycerides, and cardiovascular disease<ref name="Grundy-2012"/>. (Note that while glucose lowering drugs may help prevent the conversion of IGT to Diabetes, but it's unclear if they will help prevent some of the health complications of diabetes<ref name="Grundy-2012"/>.)
* IGT is a risk factor for liver cirrhosis<ref name="Nishida-2006"/> and survival rates for those with liver cirrhosis<ref name="García-Compeán2014"/>.
=Individuality and the Glycemic Index=
Most metrics around foods are based on relatively simple, objective measures of the food itself. In contrast, the Glycemic Index is a measure of how the human body reacts to a food, rather than just an attribute of the food in isolation. Studies have shown large variability in the Glycemic Index measured for given foods between individuals as well as large variability for repeated tests on the same individuals<ref name="Vrolix-2010"/><ref name="Vega-López-2007"/>. A 2015 study attempted to understand and model the individual Glycemic Index responses<ref name="ZeeviKorem2015"/>. This study used continuous glucose monitoring of 800 subjects for a week, with the subjects recording their food intake, sleep, and exercise electronically. This provided a remarkable 1.5 million glucose measurements over 5,435 person/days. The study found a number of things:
* A higher Glycemic Response (iAUC) was related to obesity (greater BMI), higher [https://en.wikipedia.org/wiki/Glycated_hemoglobin HbA1c ] (an indicator of average blood glucose over time), fasting blood glucose, and age. This tends to confirm the importance of Glycemic Index, but underlines the individual variability.
* Unlike earlier studies, this study indicates that the glycemic response has a lower variability for a given individual.
* Glycemic Index is normalized to the Glycemic Response for a standardized meal, which makes it a relative rather than absolute measure. This study did evaluate the absolute Glycemic Response, and found large variability for the standardized meals. Even [[Fructose]], which has lower variability, still has a few individuals with a surprisingly high absolute Glycemic Response (see below).<br/>[[File:ZeeviF3 E H.jpg|none|thumb|400px| High levels of variability in absolute Glycemic Response for standardized meals.]]
* Even when the Glycemic Responses are normalized to that of glucose, there are still remarkably high levels of variability. The diagram below shows that white bread, which is typically reported as having a Glycemic Index of around 71 has some individuals with extremely low Glycemic Index responses, and some that are much higher. Even fructose has some individuals with an indicated Glycemic Index of over 100, when the literature indicates it is generally around 19.<br/>[[File:ZeeviF3 I K.jpg|none|thumb|400px| The variability when three standardized meals are normalized to the individuals Glycemic Response for glucose.]]
* This individual variability is a nicely shown in the blood glucose responses shown below. These graphs show the responses of two individuals to glucose and bread. Each test was repeated, giving two lines for each participant's response to each of the two foods. You can see that each participant had reasonably similar responses each time they ate each of the foods. However, the top participant had a much higher Glycemic Response to glucose than a bread, this is what would be expected from the published Glycemic Index values. However, the participant in the lower graph has an inverted response, with higher glycemic response to bread than glucose.<br/>[[File:ZeeviF2 E.jpg|none|thumb|300px| The blood glucose responses to the standardized meals of glucose and bread for two individuals.]]
* The cloths below show two individuals Glycemic Response to an honors and cookies. Here you can see the individuals had opposite responses to the two foods. Personally, I am a little cautious in interpreting this data as the honors and cookies were not standardized. It's possible that the top participant had an extremely ripe banana, and a high fat cookie, while the lower participant had an unripe banana, and a low fat cookie.<br/>[[File:ZeeviF2 G.jpg |none|thumb|300px| The blood glucose response to anonymous and cookies in two individuals.]]
* While the study found large variability, it also found that the average Glycemic Response generally corresponded well to the published figures.<br/>[[File:ZeeviF2 F.jpg|none|thumb|300px| The average Glycemic Response in absolute terms to various foods.]]
The research team used the data from the 800 participants to create a computer model to predict the Glycemic Response of a further hundred subjects. This computer model was able to predict the Glycemic Response surprisingly well (r=0.70 for those into statistics.) Analysis of the computer model revealed a number of factors that tend to predict the glycemic response.
* As expected, the carbohydrate content of a meal is strongly related to the Glycemic Response. However, the model did reveal some subjects that seemed relatively insensitive to the overall carbohydrate content. It appears that about 5% of subjects had a Glycemic Response that was as high as the rest of the subjects, but wasn't related to the overall carbohydrate content. Below you can see the relationship of participant 49 whose glycemic response is broadly proportional to the carbohydrate response compared with participant 145 whose glycemic response appears unrelated to the overall carbohydrate content. <br/>[[File:ZeeviF4 B.jpg|none|thumb|300px| The glycemic response of two individuals to meals with varying levels of carbohydrate.]]
* The study also found the expected relationship between glycemic response and fat content. However, there were also individuals where the relationship between fat content glycemic response was not seen.
* The level of Fiber in a meal was linked to higher glycemic responses. While this seems counterintuitive, I suspect it's because meals that are high in fiber are often also high in carbohydrates, and finely ground whole wheat products often exhibit similar glycemic index is to those using white wheat flour. Higher fiber intake over a 24-hour period was related to lower glycemic responses.
* There were also indications that higher sodium levels, longer times since the last sleep and the subjects' cholesterol levels were also linked to higher glycemic responses.
* The study analyzed the digestive bacteria from stool samples, and found 21 beneficial strains and 28 non-beneficial strains of bacteria.
All this suggests that while the published Glycemic Index values are a useful rough guide to the health implications of various foods, they don't show the whole picture. It is possible to measure your own glycemic response, and I have some recommended [[Blood Glucose| Blood Glucose Meters]]. While it is impractical to perform this type of test on a wide variety of foods, I've measured my own glycemic response to a few of the meals that I eat most commonly. This measurement must be performed first thing in the morning on an empty stomach. Simply measure your fasting blood glucose, consumer the meal, then check your blood glucose over the next two hours. Measuring at 15, 30, 45, 60, 90, and 120 minutes will give you a standardized curve, but you could probably get away with far fewer measurements to get an approximation. You could normalize your values to a standardized meal, but I simply used an absolute value so that I could confirm what I was eating was not creating a health problem for me.
=Simple and Complex Carbohydrates=
At one time, it was believed that "simple carbohydrates" had high Glycemic Index, while "complex carbohydrates" had lower Glycemic Indexes<ref name="ND"/>. The difference between simple and complex carbohydrates is based on the chemistry of the carbohydrate molecule, with small molecules like sugar considered "simple" and big molecules like bread considered "complex". This division into simple and complex is unfortunately crap (biochemistry term meaning 'not useful'). The digestion of carbs is a sophisticated system that does not follow this simple division. Some simple carbs (Fructose) are very slow to digest, whereas some complex carbs (maltodextrin) are very easy to digest. For instance, white bread (a "complex" carb, GI 70) has a higher Glycemic Index than table sugar (a 'simple' carb, GI 60). This is because highly refined flour in bread is more easily digested than table sugar (which is half fructose).
=References=
<references>
<ref name="ZeeviKorem2015">David Zeevi, Tal Korem, Niv Zmora, David Israeli, Daphna Rothschild, Adina Weinberger, Orly Ben-Yacov, Dar Lador, Tali Avnit-Sagi, Maya Lotan-Pompan, Jotham Suez, Jemal Ali Mahdi, Elad Matot, Gal Malka, Noa Kosower, Michal Rein, Gili Zilberman-Schapira, Lenka Dohnalová, Meirav Pevsner-Fischer, Rony Bikovsky, Zamir Halpern, Eran Elinav, Eran Segal, Personalized Nutrition by Prediction of Glycemic Responses, Cell, volume 163, issue 5, 2015, pages 1079–1094, ISSN [http://www.worldcat.org/issn/00928674 00928674], doi [http://dx.doi.org/10.1016/j.cell.2015.11.001 10.1016/j.cell.2015.11.001]</ref>
<ref name="Brand-MillerStockmann2008">J. C Brand-Miller, K. Stockmann, F. Atkinson, P. Petocz, G. Denyer, Glycemic index, postprandial glycemia, and the shape of the curve in healthy subjects: analysis of a database of more than 1000 foods, American Journal of Clinical Nutrition, volume 89, issue 1, 2008, pages 97–105, ISSN [http://www.worldcat.org/issn/0002-9165 0002-9165], doi [http://dx.doi.org/10.3945/ajcn.2008.26354 10.3945/ajcn.2008.26354]</ref>
<ref name="ND">http://nutritiondata.self.com/topics/glycemic-index, http://nutritiondata.self.com/topics/glycemic-index, Accessed on 13 February 2016</ref>
<ref name="Ludwig-1999">DS. Ludwig, JA. Majzoub, A. Al-Zahrani, GE. Dallal, I. Blanco, SB. Roberts, High glycemic index foods, overeating, and obesity., Pediatrics, volume 103, issue 3, pages E26, Mar 1999, PMID [http://www.ncbi.nlm.nih.gov/pubmed/10049982 10049982]</ref>
<ref name="MastersReither2013">Ryan K. Masters, Eric N. Reither, Daniel A. Powers, Y. Claire Yang, Andrew E. Burger, Bruce G. Link, The Impact of Obesity on US Mortality Levels: The Importance of Age and Cohort Factors in Population Estimates, American Journal of Public Health, volume 103, issue 10, 2013, pages 1895–1901, ISSN [http://www.worldcat.org/issn/0090-0036 0090-0036], doi [http://dx.doi.org/10.2105/AJPH.2013.301379 10.2105/AJPH.2013.301379]</ref>
<ref name="Vega-López-2007">S. Vega-López, LM. Ausman, JL. Griffith, AH. Lichtenstein, Interindividual variability and intra-individual reproducibility of glycemic index values for commercial white bread., Diabetes Care, volume 30, issue 6, pages 1412-7, Jun 2007, doi [http://dx.doi.org/10.2337/dc06-1598 10.2337/dc06-1598], PMID [http://www.ncbi.nlm.nih.gov/pubmed/17384339 17384339]</ref>
<ref name="Vrolix-2010">R. Vrolix, RP. Mensink, Variability of the glycemic response to single food products in healthy subjects., Contemp Clin Trials, volume 31, issue 1, pages 5-11, Jan 2010, doi [http://dx.doi.org/10.1016/j.cct.2009.08.001 10.1016/j.cct.2009.08.001], PMID [http://www.ncbi.nlm.nih.gov/pubmed/19737630 19737630]</ref>
</references>