Smart Bytes®

Feeling nutrition info overload? I will help you sort through to find what’s important to you. Read more. . .

Subscribe to Blog

FREE Gift: Easy Nutrition Upgrades


Carbs, Glycemic Index & Cancer Risk: New Studies in Perspective

New studies are adding to the picture of how our eating habits, including the carbohydrate-containing foods we choose, affect our cancer risk.  Although the amount of data is building, the picture is still anything but clear when we stack the studies next to one another.Carbs, GI and Cancer Risk - New Studies

This week I’m in Indianapolis, where the American Association of Diabetes Educators asked me to speak at their annual meeting. My topic: At the Crossroads: How do Diabetes and Cancer Intersect? As I’ve talked with colleagues here, the question of whether glycemic index is part of this intersection has come up several times.

Let’s step back from the relatively large volume of studies focused on glycemic index and cancer, and look at how we can weave together the common threads into a picture that makes sense.

Glycemic Index vs. Glycemic Load

Carbohydrate that is digested and absorbed quickly, raising blood sugar rapidly and leading to a prompt rise in insulin to handle it, is referred to as having a high glycemic index (GI).  Such foods include not only sweets, but potatoes, white rice and refined cereals and breads.  Foods with a low GI are digested and absorbed more slowly. These include whole grain breads, pasta, beans, and most fruits and vegetables.

The concept is simple, but, as they say, “the devil is in the details”….

  • Although you can find lists of foods with their GI value in books and on websites, actual blood sugar-raising effects of foods vary with how they are prepared.  So depending on the reference used, the GI listed for a food may or may not represent the effect of the food as you are eating it.
  • What’s more, the actual blood sugar effects of any food change substantially based on whether you eat it on its own or along with sources of protein, fiber or fat, all of which blunt the rise in blood sugar.
  • Furthermore, a large portion of a “low GI food” could end up raising blood sugar as much as a small portion of a “high GI food”. This is where Glycemic Load (GL) comes in, which is calculated based on the glycemic index of a food and its portion.  In other words, GL is an expression of both type and amount of carbohydrate in a food or in an overall diet.

Why Cancer?

High levels of insulin circulating through the body seem to create metabolic conditions that favor cancer development.  Insulin is not just a blood sugar control hormone. It also activates pathways that lead to growth and reproduction of cells, including any cancer or pre-cancerous cells present. In addition, high levels of insulin and a related growth factor (IGF-1) seem to inhibit the self-destruction of abnormal cells that is part of our body’s normal defense against cancer.  What’s more, high levels of insulin lead to changes in proteins that tie up estrogen, resulting in greater availability of estrogen to promote estrogen-sensitive cancers.

The Latest Studies

The large EPIC study, a population study based on 23 centers throughout Europe, recently released analysis looking at breast cancer risk’s potential link to GI and GL. It showed no link of GI or GL to overall breast cancer risk or to estrogen-sensitive (ER-positive) postmenopausal breast cancer, which is the most common form. However, women with overall eating patterns with high GL did show 36 percent greater risk of ER-negative postmenopausal breast cancer compared to women with the very lowest GL. Furthermore, women with the very highest total carbohydrate consumption showed a similar 41 percent greater risk of this form of breast cancer compared to women eating the least. This is interesting follow-up to the very mixed bag of results from numerous previous studies of breast cancer. The most recent analysis combining multiple studies together does link the very highest GI with a statistically significant but modest 8 percent increased risk of breast cancer compared to those with lowest GI, but individual studies of GI and GL have shown inconsistent results.  So this recent study may be the beginning of finding the specifics of which forms of this cancer and in whom GI/GL might be relevant.

The other hot-off-the-press study, a Canadian population study, continues the picture of mixed results. Here there was no link to GI or GL seen for pre- or postmenopausal breast cancer (no analysis by estrogen receptor type) nor for 11 other cancers.  However, prostate cancer incidence was 26 percent higher in men with diets highest in GI. And both colorectal and pancreatic cancer incidence was higher (28 and 41 percent, respectively) in people with high-GL diets.  Surprisingly, since the theory would be that increased cancer risk reflects the effects of elevated insulin levels, the effect of GL on cancer was no greater among those with most overweight and least physical activity – which is the group of people most likely to be “insulin resistant” and respond to high-GL diets with increased insulin levels. Also surprisingly, looking at breast cancer results the trend was not statistically significant (which means it can just be a matter of chance), but if anything, the trend was for diets highest in GL to be linked to lower incidence of postmenopausal breast cancer.

These results showing a link to colorectal cancer add to the picture, but they are not the whole picture.  An analysis of 14 population cohort studies (the strongest type of population research) published just a few months ago showed no link to colorectal cancer risk related to total carbohydrate consumption or GL.  There was a trend for a seven percent increase in risk among those with the very highest GI, but this was not statistically significant, which means it could be found simply by chance.

It’s Got to be the Details

The wide variation in results of the many studies that have now looked at the potential relationship of GI and GL to cancer risk show that this relationship is not clear-cut.

  • Some variation can come from differences in the studies themselves, for example differences in the forms used to ask about people’s diets and the tools used to analyze GI. Even when the GI score of a food is based on detailed description of how it is prepared, scores can’t reflect the different impact of other foods consumed at the same time.
  • Studies differ in what other influences they adjust for in analysis of GI/GL and cancer risk.  For example, the latest Canadian study reported above did not control for dietary fiber consumption, whereas the EPIC study of breast cancer risk did.  Dietary fiber in a food or in other foods eaten at the same time reduces the blood-sugar rise it produces.  In part, dietary fiber of a particular food is already reflected in its GI, but overall dietary fiber consumption could still affect its impact.  Analysis by the American Institute for Cancer Research/World Cancer Research Fund in the continuous update project looking at all available research worldwide links higher dietary fiber with lower risk of both breast and colorectal cancers.
  • Individual genetic and metabolic differences may make some people respond to similar food choices with varying jumps in insulin levels.
  • Since insulin does not act alone in promoting cancer, other individual differences may make similar increases in insulin weigh more or less heavily among the many influences promoting and inhibiting cancer development.

Bottom line: We eat food, not an index

New American Plate shows how healthy carbs fit in a cancer-preventing diet

The New American Plate by AICR

Evidence is now quite strong that high levels of insulin lead to metabolic conditions favoring cancer development.  Awareness of glycemic index can help, but it can also lead you astray.  Just because two foods are similar in glycemic index does not mean their total impact on health is the same.

Rather than focus specifically on the glycemic index of your diet, aim for an overall strategy to avoid elevated insulin and create a pattern of eating choices that supply nutrients and phytochemicals that reduce cancer risk.

♦ Accumulate at least 30 minutes of moderate physical activity a day to decrease insulin resistance. Don’t let more than a day go by without being active.

       ♦ Control portion sizes even of “healthy” food.  This has immediate benefits on GL, but even more importantly is key to reach and maintain a healthy level of body fat, which acts in multiple ways to reduce cancer risk.

       ♦ Make vegetables, fruits, whole grains and beans the largest part of your plate.  They’re more than just low GI: they provide dietary fiber that seems to have protective effects beyond its impact on blood sugar and thousands of phytochemicals that act as antioxidants and apparently with more direct effects on cells to decrease cancer development.


For the opposite of the complexities of glycemic index, check the very simple approach to balanced eating and sensible portions illustrated in the New American Plate developed by the  American Institute for Cancer Research (AICR).


Romieu I et al. Dietary glycemic index and glycemic load and breast cancer risk in the European Prospective Investigation into Cancer and Nutrition (EPIC). Am J Clin Nutr 2012.  96:345–55.

Dong JY, Qin LQ.  Dietary glycemic index, glycemic load, and risk of breast cancer: meta-analysis of prospective cohort studies. Breast Cancer Res Treat 2011;126:287–94.

Hu J et al. Glycemic index, glycemic load and cancer risk. Annals of Oncology. Published online July 25, 2012, ahead of print.

Aune D et al. Carbohydrates, glycemic index, glycemic load, and colorectal cancer risk: a systematic review and meta-analysis of cohort studies. Cancer Causes Control. 2012. 23(4):521-35.

Aune D et al. Dietary fiber and breast cancer risk: a systematic review and meta-analysis of prospective studies. Ann Oncol. 2012.  23 (6): 1394-1402.

Continuous Update Project Report Summary. Food, Nutrition, Physical Activity and the Prevention of Colorectal Cancer. World Cancer Research Fund / American Institute for Cancer Research. 2011.

4 Responses to Carbs, Glycemic Index & Cancer Risk: New Studies in Perspective

  • Dear Karen,
    Since I was unable to attend AADE this year, when I saw this post on GI/GL and cancer, I was anxious to see what I could learn from the synopsis of your talk. I’ve always found your information to hit the mark while being succinct – two great literary qualities!
    Well, this time I was disappointed for a number of reasons. My files are bulging with positive relationships between GI/GL and risks of certain cancers. I wondered why you didn’t mention any of them. Here is one at the top of a pile of articles: “Glycemic index, glycemic load and cancer risk,” published In Annals of Oncology, June, 2012 by J. Hu et al. The researchers found that “Dietary GI was positively associated with the risk of prostate cancer. A higher GL significantly increased the risk of colorectal, rectal and pancreatic cancers.” As I say, this is one of many. To be sure, some studies failed to show beneficial results, but I think you would find that these are fewer in number than the positive ones. I question why a broader (more than 2 studies) and more balanced presentation of the facts did not occur.
    I also would like to mention some inaccuracies in the generalizations you have made about high/low GI foods. You state that sweets are high GI. Not all sweets are high GI. Peanut m&ms, for instance, have a GI of 33 (low); low fat vanilla ice cream has a GI of 36 (low). You state that potatoes, rice and bread are high GI – but not all of them – the smaller new potatoes, several varieties of Uncle Ben’s rice and some rye, pumpernickel and sourdough breads – these all have low/moderate GI values. You also mention pasta as a low GI carb. This is true if it is not over-gelatinized by over-cooking. You are correct that preparation impacts the rate of digestibility of a carb. And if we eat spaghetti that is al dente or over-cooked, our body will have to deal with what we swallow. The GI is only telling us what the body already knows! Please do not believe that all dietary fiber slows down gastric emptying and postprandial glycemic excursions. This is only true of soluble fiber. You’ll find all of this information in the research. I also think it is important to distinguish between a low GI “food” and a low GI carbohydrate-rich food. Beef and eggs and oils are, technically, low GI “foods” but they have nothing to do with the glycemic index.
    I have come to know about the glycemic index from having carefully followed the literature for the past 20 years, visited the premier GI testing lab in Sydney, Australia, published research on the use of the glycemic index in my MNT practice for the past 20 years. I don’t consider myself an expert but it is clear to me when biased misinformation is presented. It is an injustice to all concerned.
    Johanna Burani MS,RD,CDE

    • Karen says:

      Thank you so much, Johanna, for taking time to read and comment on my blog post about the two recent studies on glycemic index and cancer risk. I’m not sure if I was unclear in making the points I was trying to convey, or if this is simply an instance of professionals with different interpretations of currently completed research, so I will try to clarify.

      First, my goal in writing this post was to address two specific recent studies on glycemic index (GI) and cancer risk that have been getting media attention and put them into the big picture of overall research. One study was the EPIC study on breast cancer. And I noted that by looking separately at ER+ and ER- cancer, these researchers found a link between GI and ER- cancers that might have been obscured by previous work that looked at all breast cancer together. The second new study was the Canadian population study, which one can see either by clicking on the link or by checking the reference list at the bottom of the post, is exactly the study by Hu that you suggest I read. I specifically noted the 26% higher prostate cancer risk in men with diets highest in GI that you point out. I also noted this study’s link of highest GL diets with increased colorectal and pancreatic cancer risk; and the lack of observed link to breast cancer when viewed overall (rather than by ER subtype as in the EPIC study). Due to space concerns, I did not discuss the lack of observed association between GI or GL and several other cancers, nor the trend (not statistically significant) for higher GI to be associated with reduced risk of liver cancer. The latter would be quite unexpected based on the link between liver cancer and type 2 diabetes, given the assumption that elevated insulin levels are associated with the cancer-diabetes connection.

      After noting the results of these two studies, I simply tried to address the question of how they fit into the overall picture of research at this time. As we all know, one can pull out studies to back up any point of view, especially on a topic like this that has had a wide mixture of results through the years. Therefore, for the sake of the succinctness that you identify as something I aim for, I used the most recent meta-analyses that address glycemic index and risk of these cancers as helpful tools to identify the current overall state of this research. And that answer, at least as I see it, is still cloudy. I am not saying that GI and GL are not factors in cancer risk. However, other aspects of our eating habits, as well as our metabolic and genetic differences, are making it hard to pin this down. For example, I note that although the EPIC study found a link with ER- breast cancer even after adjusting for dietary fiber intake, Hu et al’s findings linking GI with prostate cancer and colorectal cancer did not include any adjustment for fiber. That opens a whole different discussion: are the viscous types of soluble fiber part of what creates a low GI and thus shouldn’t be adjusted for in these studies, or is the question whether GI has an effect beyond fiber intake, and thus should be adjusted? I was just pointing out differences in methodology of the studies that one needs to consider when interpreting the results.

      Finally — and kudos to anyone who is still reading at this point! — my bottom line question was not meant to address the question of research in this field, but the question of what we tell people about their eating choices while we await clearer answers from research. My concern is exactly that which you express in your comment: as I specifically said in the post, many factors about a food’s composition and preparation affect its GI, and many of the simplistic rules or overly simplified tables of foods’ GI values mean that people trying to implement this as yet another eating guideline may not be accomplishing what they think. So the question I was trying to pose was whether, given the guidelines we already have about steps that have strong research behind them about how to reduce cancer risk (including that related to the insulin resistance and inflammation of type 2 diabetes), people should be also trying to adjust the GI or GL of their diets.

      In your study of your work with patients with diabetes ( ), you showed that the 21 people in the study whom you counseled clearly benefited from this approach. I think we have to be careful about assuming results of people working with someone like you, skilled in working with GI, would necessarily be the same as someone just trying their best to take information they find about GI and piece it together with other nutrition information. It would be very interesting to see the results of a randomized controlled trial in which we could compare the results of your work counseling those patients with outcomes of equally intense expert counseling using another approach. In some studies of weight loss, for example, simply having access to good nutrition counseling made more of an impact than any differences in content of the counseling. And please note that though in your study you were looking at how GI impacted weight and diabetes control, the outcome being discussed in my blog post and in these two studies was the impact on cancer risk. It’s possible that GI and GL could have important effects on blood sugar control and yet not add a significant piece to the question of cancer risk beyond what is already covered in current guidelines to lower cancer risk.

      In fact, my presentation at the conference that you mention was really not about glycemic index and cancer risk. It was about the multiple ways that risk of type 2 diabetes and cancer risk intersect. My bottom line point was that for people with diabetes, the evidence now suggests that for their overall health (including cancer risk), it’s important that they look not only at blood sugar control, but also to other ways in which their lifestyle choices impact their health. For example, it means looking beyond plant foods as sources of carbohydrate to seeing them as sources of fiber that promotes formation of apparently-protective butyrate in the colon (acting beyond its impact on glucose absorption) and affects the microbiome of the gut, and as sources of thousands of phytochemicals that act as antioxidants and apparent epigenetic agents (changing the expression of genes such as tumor suppressor genes).

      Again, I thank you sincerely for taking the time to post a comment with your views. I very much want this blog to be a place where health professionals and the general public can discuss their take on research and how they are trying to live healthfully. I hope that my response here clarifies what I was trying to say, whether that highlights that we are more in agreement than it initially seemed to you, or confirms that we are looking at current research through different lenses at the moment.

      • Hi again, Karen,
        Thanks for taking the time to respond to my recent comment. As I mentioned, I was not at AADE to hear your talk and from your blog post it wasn’t apparent to me that you were just highlighting these two specific studies and not research in general. Thanks for that clarification. I’d like to clarify too just a few ideas from my point of view.

        Please understand, speaking of research, that I did not mention my published study in reference to my outcomes and cancer. Cancer risk and GI/GL was not my focus. I mentioned it to affirm my involvement with and attention to GI research over the past 20 years. However, in your addressing the outcomes of my study, you bring up a good point. Is it capable low GI nutrition counseling that made the difference or is it the concept itself that worked? My answer would affirm that we as counselors make a positive impact on our patients when we present solid nutrition information that is doable in their specific lifestyles long after they stop coming to MNT appointments. Some of my subjects I hadn’t seen for 3 years when I invited them to participate in the study. So my conclusion about that study in this regard is that the GI concept can and does stand alone in “the trenches” where my patients live and make their daily food choices.

        It is not clear to me if you find the GI concept too simple or too difficult. You characterize the GI as having “simplistic rules or overly simplified tables of foods’ GI values” but you also recommend deferring to the the plate method (not mutually exclusive with GI and an excellent teaching tool in my opinion) “for the opposite of the complexities of GI.” You also noted that “GI can lead you astray.”
        I have been teaching my patients how to implement the low GI carbs of their choice for 20 years. And I have seen how effectively it’s worked for 20 years. Is it the only approach that works? I don’t believe so. Wholesome, unrefined, nutrient-dense, well-balanced meals remain the foundation of my conversations with my patients. This mindset works seamlessly with the glycemic index concept and I haven’t found anything yet that works better. If that changes tomorrow, then my counseling will also change tomorrow.

        It’s important, and I’m sure you agree, that we must do our best to keep learning with an open mind. As we keep up with the literature we surely get more insights into our understanding and interpretation of what we learn. I will continue to keep to my practice of reading as much as I possibly can about topics of interest coming from different points of view. We are all out there everyday doing our best. Of this I have no doubt. Thanks again for a very interesting and informational exchange .

        My best regards,

        • Karen says:

          Thanks, Johanna, for the further clarifications. Clearly there is more research to come on this. I’d just say that for those who can learn to implement GI accurately and confidently from a well-trained professional such as yourself, if labs & other tests suggest benefit, then great. For those who are trying out GI-focused eating goals on their own and find it confusing, or for those who don’t see evidence of it adding any benefit beyond what is accomplished with other steps, it may be helpful to know that solid research has not yet consistently identified GI as an essential component of a strategy to lower cancer risk. It will be wonderful when research moves us further ahead in clarifying which, if any of us, particularly benefit from adding an additional goal to their healthy eating plans.
          Anybody else want to weigh in based on their experience?

Leave a Reply

Your email address will not be published. Required fields are marked *