Lifestyle+ — 17

It took a landmark 2015 paper from the scientists at the Weizmann Institute in Israel to lay the foundation for personalized nutrition. They studied eight hundred participants without diabetes in depth. This was an enormous dataset with about 1.5 million continuous glucose sensor measurements to more than five thousand standardized meals, another forty-seven thousand meals of usual food intake, and data from the participants’ gut microbiome, blood tests, and physical measurements. With a machine learning model that accounted for more than one hundred factors that could be involved in glucose spikes, the key determinant turned out to be the person’s gut microbiome, which is shaped by our diet, genetics, social contacts, and medications.

The importance of the gut microbiome for predicting our response to food was later confirmed and extended by a group in the United Kingdom using a similar big data and machine learning model approach, which added identical twins as part of the participant group, and measuring not just glucose but also triglycerides. The latter requires frequent blood test assessments since there is no wearable sensor for this key metric. Notably, there was no relationship between the glucose and triglyceride response in participants, telling us that tracking glucose alone, although helpful, is not sufficient to understand personalized nutrition. Further, it’s not just the genes, since there were marked differences among identical twins in nutrition response. While these studies highlighted the gut microbiome as a dominant contributor for our unique response to food, many other factors such as sleep patterns, physical activity, and sequence of meal ingestion play a role.

The concept of personalized nutrition was tested prospectively in a randomized trial of 225 participants with prediabetes, comparing an algorithm for predicting glucose-response diet (based on clinical and microbiome data) with a Mediterranean diet. Over six months, the algorithmically based diet led to significantly improved glucose regulation. That might help people at high risk to prevent ever manifesting type 2 diabetes, or for others already diagnosed to achieve remission. But such outcomes have yet to be assessed. In another randomized trial of a personalized nutrition program, there were reductions in blood triglycerides, body weight, waist circumference, and gut microbiome diversity.