Heidi Nelson, MD, on Whether the Future of Personalized Nutrition Lie in the Gut Microbiome
Many patients have difficulty finding a diet that is right for them, often finding themselves frustrated after seemingly doing all the right things, like counting calories and carbohydrates.
Although many current dietary approaches focus heavily on nutrition labels, a new study published in JAMA Network Open suggests that the effectiveness of one’s diet may go beyond labeling and have more to do with a plethora of individual traits, including the gut microbiome and postprandial glycemic responses to foods.1
“Since it is known that metabolism in the gut is highly influenced by microbes in food, this study was a good opportunity to assess the impact of the microbiome on the body’s responses to food,” said study author and colorectal surgeon Heidi Nelson, MD, chair of the Department of Surgery and leader of the Center for Individualized Medicine Microbiome Program at the Mayo Clinic in Rochester, Minnesota.
The study included 327 participants without diabetes from Minnesota and Florida. During the study, participants were asked to record their food intake and physical activity. Other data collected during the study included anthropometric variables, gut microbial composition, and blood glucose levels.
Dr Nelson and colleagues also assessed participants’ inter-individual diversity in postprandial glycemic responses to a diverse array of foods using a predictive model first assessed in an Israeli study performed by David Zeevi et al.2 The model was described by Zeevi et al. as “a machine learning algorithm that integrates these multi-dimensional data and accurately predicts personalized [postprandial glycemic responses].”2
With this model, Dr Nelson and colleagues were able to use the data collected in the present study to subsequently predict participants’ glycemic responses to a variety of foods, which were found to vary highly among all participants. Perhaps most notably, this approach demonstrated better overall performance in terms of controlling participants’ postprandial glycemic levels compared with current standard-of-care approaches.
“Our study is proof-of-principle and helps validate that a patient’s glycemic response to a certain food may be very specific to him or her and may be influenced by individual microbial variations, among other things,” said Dr Nelson, whose strongest research interests include colorectal cancer and the gut microbiome.
Elevated glycemic levels are known to raise one’s risk for conditions such as cardiovascular disease and diabetes, making glycemic control an important part of disease prevention. Since gut microbiome composition and glycemic responses to foods can vary highly between patients, these findings may have important future implications for personally tailored nutrition and disease prevention interventions. However, further research is still needed in this area, Dr Nelson noted.
Consultant360 spoke further with Dr Nelson about these findings and their future clinical potential.
C360: Could the model you utilized in your study eventually be applicable to a variety of conditions and disease states?
Dr Nelson: That is our hope – that the principle of obtaining numerous variables that are unique to each person, putting them into an analytic framework, and then creating predictions may have utility in the field of personalized nutrition. It is similar to personalized medicine in the sense that it involves assessing patients’ unique variables and predicting how they will respond to something instead of simply averaging responses across multiple people.
Since nutrition affects so many disease states, I believe it will play a major role in helping patients find a health state that is right for them. As a clinician, I often see patients who ask for dietary advice. In general, I encourage them to eat fruits, vegetables, greens, and protein, but advising beyond that can be extremely difficult because each patient tends to respond differently to food. One diet does not seem to work for all people, whether related to weight loss, achieving consistency in bowel function, or whatever their health goal may be.
C360: Could you discuss an example of how this model could influence personalized nutrition?
Dr Nelson: An important area that the present study has helped clarify is how a food affects each individual patient. For example, one patient may have a normal glycemic response after eating a banana and a high glycemic response after eating a cookie. However, another person may have a high glycemic response to the banana and a normal glycemic response to the cookie. In a similar scenario, one of my colleagues who tested his responses to broccoli and ice cream had a high glycemic response after eating the broccoli, but a normal glycemic response after eating the ice cream.
Both of these scenarios go against conventional thinking, but to be clear, this is not to suggest that we would recommend a patient to eat cookies or ice cream instead of bananas or broccoli. Instead, clinicians could recommend that the patient consider trying a different fruit or vegetable to see if they have a better glycemic response to it.
C360: Does your study yet have a clinical takeaway? Is the tool you used in your study ready for use in practice?
Dr Nelson: I think the answer is twofold. In terms of wellness, many patients have expressed interest in understanding their glycemic responses to food and want to know whether their individual diet is working for them when it comes to appropriate glycemic control. The knowledge from this study can help clinicians guide their patients in that sense. This study and previous studies that show we can predict an individual’s glycemic response to food is proof-of-principle that we are on the right track to advancing the field of personalized nutrition.
However, I think there is still a lot more to learn about how tools like the one in our study might be valuable in medicine. It is not a medical tool at this time, but I think the pathway towards that is clearer now. Once we take tools like these to the next level by advancing them to clinical studies, we will hopefully be able to determine their patient-level impact.
C360: What areas of further research are still needed?
Dr Nelson: Although human microbiome research is still in an early stage, available data suggest that many aspects of health relate back to the microbiome, and research is emerging on how diet affects health and how it can be personalized. Further studies will be needed to determine whether personalized nutrition could help prevent conditions like prediabetes or diabetes, but the present study is an important first step. Although food labels are very important, this tool can predict individual responses to certain foods better than a food label can.
My colleagues and I, in particular, are continuing to use microbiome research to sort through issues related to diseases like colon cancer, and certainly there is strong evidence that the microbiome may play a role in the metabolism and in colon cancer. In other words, what happens in the gut is influenced by microbes in the diet, and we are starting to see a connection between diet, microbes, metabolites created by those microbes, and polyp formation.
It is encouraging that we are beginning to understand how microbes in the gut environment may shed light on how diet and environmental factors may predispose certain patients to developing colon cancer and related diseases. This is an important field, and that is where our work is leading us. This study on personalized nutrition brings more awareness of numerous different variables that we can correlate, analyze, and use to predict a patient’s glycemic response to a certain food. This knowledge will help us start to understand potential preventive solutions.
1. Mendes-Soares H, Raveh-Sadka T, Azulay S, et al. Assessment of a personalized approach to predicting postprandial glycemic responses to food among individuals without diabetes. JAMA Netw Open. 2019;2(2):e188102. doi:10.1001/jamanetworkopen.2018.8102.
2. Zeevi D, Korem T, Zmora N, et al. Personalized nutrition by prediction of glycemic responses. Cell. 2015;163(5):1079-1094. https://doi.org/10.1016/j.cell.2015.11.001.