Why Nutrition Isn’t One Size Fits All
In a Nutshell
A Stanford study reveals extreme individual variability in blood glucose responses to identical carbs—grapes spike most due to sugar, rice spikes half the participants (especially Asians) more than ice cream, potatoes/pasta hit muscle-insulin resistant subtypes hardest—enabling personalized nutrition via subtype matching and DIY tests like potato-to-grapes ratios with CGMs. Mitigators like protein/fiber blunt spikes in insulin-sensitive people but fail in resistant ones; fat only delays. Future AI tools will personalize fiber/microbiome interventions amid challenges like inflammatory reactions to kale or inulin, emphasizing trial-and-error over one-size-fits-all diets.
These notes were generated by AI and may contain inaccuracies.
Professor Michael Snyder, chair of genetics at Stanford School of Medicine and director of Stanford Center for Genomic and Personalized Medicine, discusses a Nature Medicine 2025 study: "Individual variation in glycemic response to carbohydrates and underlying metabolic physiology." The study builds on prior work subclassifying diabetes and prediabetes subtypes: muscle resistant, beta cell defects, incretin (GLP-1), and hepatic (liver insulin resistance).
In the study, 55 people ate the exact same amount of carbohydrates in different forms: grapes, bread, potatoes, pasta, rice, beans, berries. Later, mitigators (protein, fiber, fat) were tested before spiking carbs.
Goal: Match glucose dysregulation subtypes with foods causing spikes for personalized nutrition. People respond differently to foods; avoid pro-inflammatory or glucose-spiking foods.
Figure 1D heatmap showed grapes causing acute rapid glucose spikes for most people due to high sugar content. Spike is high but transient; area under curve not as bad as other carbs.
Grapes are sugar. That's really what it comes down to.
Rice was the number one spiking food for ~half the people, often prolonged. Asians spiked more to pure white rice than non-Asians. Responses varied: some to bread, potatoes, pasta; fewer to beans and berries (more fiber, carbs not in raw glucose/sugar form).
Rice worse than ice cream for many (referencing Aaron Seagal's TED talk and cookie vs. banana study showing opposite responses).
White rice has 8% protein; ancient forms like Rocky Mountain barley/oats have 20% protein, lower glycemic index.
Hypertension correlated with bread spiking (possibly due to wheat/gluten inflammation). Microbes correlated with glucose spiking.
Subtype-specific spikes:
- Muscle resistant: spike to potatoes, pasta (common: resistant starch).
- Beta cell defects (Snyder's subtype): spike to potatoes.
Potato-to-grapes ratio as assay for subphenotype: muscle resistant have higher ratio (grapes ~pure sugar baseline; potatoes spike more if resistant). Could be DIY test with continuous glucose monitor (CGM), though glucose shot more precise. Ongoing study on potato types: French fries, mashed, baked.
Healthiest potato: baked with skin on (less glycemic than others). Raw potato less glycemic (poor cellulose breakdown), though possibly toxic in high doses (anecdote: host ate raw potatoes at survival camp, vomited). Temperature affects glycemic impact. Snyder adds butter; fat delays but trade-offs (e.g., French fries fatty, may raise cholesterol).
No strict threshold for insulin resistance (continuous variable); translatable to gold standard tests. Plan to validate and release free algorithm (prefer glucose version; smaller dose for diabetics).
Everyone spiked to white rice. Preceding 10g each:
- Protein: egg white
- Fiber: chickpea (inulin major component)
- Fat: cream
Protein and fiber mitigated spikes in insulin sensitive people, but diminished in insulin resistant. Fat delayed but didn't reduce peak/area much. Theories: lines gut, but less effective in resistant.
Inulin study: improves insulin sensitivity, fatty liver; reshapes microbiome to catabolize fructose into amino acids/glutathione (immunization against sugar).
Fiber study (18 people, crossover): arabinoxylan (psyllium husk/Metamucil) decreased cholesterol significantly; inulin no effect. Individual differences: some responded oppositely. Specific microbiome shifts.
Microbiome personal, set first 3 years of life (Baylor Group work), hard to shift later. Complex guilds determine reactions. Acacia lowers HbA1c 1%. Ongoing: 4 fibers (arabinoxylan, inulin, resistant starch, beta glucan) with microbiome/glucose analysis for AI personalization. Need long-term effects data.
Fiber not just mechanical (gastric emptying); diverse interactions. Inulin caused inflammation at 30g/day in 3 people (Snyder subclinical CRP rise; others symptomatic).
70% immune cells in gut; interacts with microbes/foods. Host anecdote: early antibiotics caused ulcerative colitis; fiber (e.g., cauliflower) causes bloating. Heterogeneity challenges personalization.
Snyder: Homeostatic systems shifted early (fever/antibiotics). Test deep micronutrients (Genova, etc.) for deficiencies. Vitamin D3 supplementation correlates with health benefits.
Need general + specific knowledge for personalization.
People reacted differently to the Ensure shake: for some it was pro-inflammatory, for others anti-inflammatory. Same shake. Ensure shakes are sugar bombs, loaded with sugar—some have 50 grams of added sugar. They're stocked in hospitals and freezers throughout med school.
Speaker has been on a keto diet for 6.5 years. Doesn't mind the food, but social elements create difficulty. Views it as therapeutically necessary for the foreseeable future. Wouldn't mind never eating bread again and dying at 150.
Optimistic for the future because in 10 years we'll have an order of magnitude better understanding.
Believes technology and scientific understanding will evolve, providing more tools to try unimaginable things.
Watch for inflammation response. Concern: missing microbes not being fed by the diet.
No reliable, reputable microbiome assays yet to identify what's missing. Assessing microbiome functionality is hard because different bugs have different functions in different people. Not there yet.
Microbiome is a guild: many ways to make a good microbiome product. Hundreds of strains in the gut, some dominate. Different microbes bring different hydrolases to break down fibers. Multiple solutions to the same problem of breaking down fibers, plus personal differences, probably through immune reactions to certain fibers.
Most commercial fibers are mixtures, hard to avoid specific immuno-reactive ones empirically. Figure out what works by trial and error.
Huge glycemic response to kale, more than milk chocolate. Not much sugar in kale. Likely inflammatory reaction spiking sugar, possibly via cortisol. Not sugar release from kale digestion, but immune reaction causing secondary glycemic spike.
Wonder about specific fibers in kale that trigger reaction, with hydrolases ready to attack efficiently.
Hard to predict how different bodies respond. Major takeaway: individual variability in responses.
Future goal: profile at baseline to identify reactions and beneficial foods, setting personalized goals.
Speaker has done N=1 experiments published on self. Offers as deep profiling subject, willing to fly to Bay Area. Lab successful in processing blood samples.
Thanks for fourth interview, invitation for fifth and sixth. Enjoy conversations.
Check work at Snyder Lab website—studies running there. Search Snyder on YouTube and podcasts.
Keep Nick Norwitz in your library
Save the videos and channels worth coming back to, and find them again in one place.
